3GPP TSG RAN WG1 Meeting #120bis	R1-2501708
Wuhan, China, April 7– April 11, 2025
Agenda Item:	9.4.3
Source:	Futurewei
Title:	Discussion on timing acquisition and synchronization aspects for A-IoT
Document for:	Discussion and Decision 

Conclusion
Proposal 1: Adopt at least Option 1 Alt1 for SIP design, with a single ON-OFF transmission, i.e., one high-voltage transmission followed by one low-voltage transmission, is used. The durations of ON and OFF may be the same or different.

Proposal 2: Adopt at least Option 1 Alt1 for clock acquisition design, where the length is changing with the M values.
Proposal 3: Study to use circular shifting method to remove CP issue in TAS.  
Proposal 4: For D2R transmission, D2R preamble +  midamble(s), where  is supported. 
Proposal 5: For D2R transmission, preamble and midamble signals support M-sequences. 

Observation 1: Among the 4 SIP designs (Alt 1-1, Alt 1-2, Alt 1-3, and Alt 2-2), Alt 1-2 and Alt 2-2 provides better performance

Observation 2: CP issues could be avoided if SIP and CAP are within one OFDM symbol.
Observation 3: The device determines the end of R2D transmission by detecting patterns that violate Manchester coding rule.
Observation 4: It is benefit to the device if the pattern at end of PRDCH transmission is known to the device.
Observation 5: Golay and M-sequences show similar performance. 
Observation 6: Reader can explicitly indicate the number of D2R midamble fields in the R2D transmission or implicitly indicated by which the device determines the number of D2R midamble fields using, e.g., the length of the midamble sequence as a reference.    
Observation 7: For D2R transmission, the overhead incurred by preamble and midamble sequences can be reduced if such sequences are used to carry one or more information bits.   

3GPP TSG-RAN WG1 Meeting #120bis	R1- 2501721
Wuhan, China, April 7th – April 11th, 2025

Agenda Item:	9.4.3
Source:	Ericsson
Title:	Timing acquisition and synchronization for Ambient IoT
Document for:	Discussion, Decision
1	

Conclusion
In the previous sections we made the following observations:
Observation 1	Golay sequences offer perfect autocorrelation when used in complementary pairs, making them ideal for accurate channel estimation and synchronization in multipath conditions, though they require more complex processing.
Observation 2	D2R preamble length should be designed to strike a balance between detection/estimation performance and overheads due to signaling and transmit power consumption.
Observation 3	The simulated D2R preamble can tolerate SFO up to 1% (AWGN) with up to 1 dB loss in performance for a sufficiently long preamble sequence length (32 or greater).
Observation 4	The simulated D2R preamble cannot tolerate SFO beyond 1% (AWGN) without a significant loss in performance, which suggests that it is sufficient to target a residual SFO of 1% or lower after SFO correction.
Based on the discussion in the previous sections we propose the following:
Proposal 1	For SIP of the R-TAS consider to down select between Alt 2-2 or Alt 2-3 with patterns of ON-OFF-ON or OFF-ON-OFF where each part has the duration of (1/6,1/6,4/6) of an OFDM symbol, respectively.
Proposal 2	For the clock-acquisition part, option 2 can be adopted, where the duration of clock-acquisition part is fixed, and higher values of M utilize a higher repetition factor.
Proposal 3	For the clock-acquisition part, consider the overall duration of 2 OFDM symbols, depending at the end stat of the SIP, the base of CAP signal can be ON-OFF-ON or OFF-ON-OFF in each OFDM symbol. The CAP signal can be a function of “M”, by repeating a reference signal as following (as in Table 2):
at the OFF patterns if the base is ON-OFF-ON.
at the ON patterns if the base is OFF-ON-OFF.
Proposal 4	The TBS included in the R2D control information to explicitly indicate the end of PRDCH transmission. However, in combination with this, a simple postamble with a distinct pattern—distinguishable from SIP, and PRDCH—can be used, especially in cases with high SFO.
Proposal 5	The complementary pairs of Golay sequence can be used as D2R preamble .
FFS: down selecting between the following two lengths:
Alt 1: two sequences of length 16.
Alt 2: two sequences of length 32.
Proposal 6	In case of D2R midamble, the presence, location, and number of D2R midambles is based on explicit indication by the reader. FFS: condition for midamble presence and indication either by L1 control information or L2 control information.
Proposal 7	The exact number of midambles to be added to a PDRCH transmission can depend on the following parameters: (i) the transmission length/duration, (ii) residual timing error, (iii) whether coherent detection is used at the reader, (iv) channel and interference conditions and estimation needs.

3GPP TSG RAN WG1 #120b                                            R1-2501735                
Wuhan, China, April 7th – April 11st, 2025

Source:	TCL
Title:	Discussion on timing acquisition and synchronization functionalities for Ambient IoT
Agenda Item:	9.4.3
Document for:	Discussion and Decision

Conclusion
In this contribution, we provide our views on the frame structure and required functionalities of R2D/D2R signal for AIoT. The observations and proposals are listed as below:
Observation 1: If the start of SIP is not aligned with that of OFDM symbol, and M value is small (e.g., 4), single ON-OFF pattern for SIP could be considered.

Observation 2: If the start of SIP aligned with that of OFDM symbol, and M value is relatively large (e.g., 16), multiple ON-OFF pattern for SIP could be considered.

Observation 3: To ensure the detection and activation performance at device 1 side, first ON should be long enough, e.g., more than 50% length of entire start-indicator length.

Observation 4: For Alt 2-1 of multiple ON-OFF transmission, it could be achieved by repetition based single ON-OFF transmission.

Observation 5: In SI phase, one agreement has given for clarifying how to design basic SIP pattern, i.e., one high-voltage transmission followed by one low-voltage transmission, where ON and OFF may have same or different duration.

Observation 6: Device detects one fixed length of SIP (e.g., one OFDM symbol) with different R2D data rate, which might be able to reduce the detection complexity and power consumption.

Observation 7: If start-indicator with M=32 during one OFDM symbol and ON-OFF pattern is detected by device, device will not be waked-up due to the limited duration and energy when SCS equals to 15KHz

Observation 8: Before discussing 1/2 OFDM symbol as entire length of SIP, it is more clear to give some conclusions including whether the start of SIP needs to be aligned with that of OFDM symbol and minimum M value.

Observation 9: If edge/transition method is used for start-indicator detection, the first level (e.g., ON level) will not be detected by device but it can activate/wake-up device to receive following R2D signal, e.g., clock-acquisition. 

Observation 10: If the last level of start-indicator is OFF and CP of CAP is ON, transient period from OFF to ON should be discussed to avoid the CP recognition, for example, transient period for raising edge should be less than the length of CP of CAP

Observation 11: In RFID, the length of R=>T calibration in preamble is [2.5Tari~3Tari] (one tari value or tari length is regulated in the range of 6.25 µs to 25 µs), which is longer than delimeter (star-indicator).

Observation 12: For different CAP pattern, there is different methods to calculate R2D chip duration 
For ON-OFF-ON-OFF pattern of CAP, chip duration equals to T1/2.
For ON-OFF-OFF-ON pattern of CAP, 
If OF-ON following ON-OFF-OFF-ON, i.e., ON-OFF-OFF-ON-OFF-ON, chip duration equals to T1/3.
If ON-OFF following ON-OFF-OFF-ON, i.e., ON-OFF-OFF-ON-ON-OFF, chip duration equals to T1/4.

Observation 13: One simpler method to calculate R2D chip duration is no need to regulate same direction of transition edge, then chip duration always equals to T1/2.

Observation 14: If CP inserted and transition edge occurs only at the start or only at the end of the CP, and subsequent chip with different level of CP, CP could be used for calculating chip duration and there is no special consideration for this case. 

Observation 15: If CP inserted and transition edge occurs during CP transmission, CP might be still used for calculating chip duration. Even CP has transition edge ON-OFF, the transition edge of end of CP (e.g., OFF-ON) could be used for calculating chip duration by combining subsequent transition edge of OFF-ON. 

Observation 16: Hopping could be achieved by change BLF or small frequency shift for D2R preamble sequence.

Observation 17: If no D2R line code is used, by using a square-wave corresponding to the small frequency-shift, the time duration Tb corresponding to each information bit includes R number of square wave periods generated by 2R OOK chips [0, 1, 0, 1 …]/[1, 0, 1, 0 ..].
Observation 18: If no D2R line code is used, by using a square-wave corresponding to the small frequency-shift, the time duration Tb corresponding to each information bit includes R number of square wave periods generated by 2R BPSK chips [-1, +1, -1, +1, …]/[+1, -1, +1, -1, …].

Observation 19: M sequence has good self-correlation performance with higher main lobe and fixed peak side-lobes, and it is better method for device 1 due to its low generation complexity. 

Observation 20: Due to maximum D2R TBS is 1000bit, performance of anti-interference might not be good than expected for Golay sequence.

Observation 21: If Mag 1 only transmits RN16, option 1 only with preamble may be enough even CRC is added. 

Observation 22: If Msg 1 transmits RN16 with data or Msg 3 transmits data, option 1 only with preamble may not be enough due to long frame structure needs Midamble to keep sync and estimate SFO.

Observation 23: For random access, TDMA and FDMA could be used for D2R transmission, including Msg 1 and Msg 3.

Observation 24: Reader may need to receive Msg 1/Msg 3 with preamble/midamble/postamble to estimate channel or interference or SFO. However, reader do not need detect or correlate all Preamble/Midamble/Postamble from each Msg 1/Msg 3 if configured. 

Observation 25: gNB needs to estimate CLI before receiving D2R signal to avoid the impact of CLI on D2R signal demodulation. When interference channel get change, gNB needs to estimate CLI again during receiving D2R signal.

Observation 26: Reader do not need to receive D2R signal when estimating CLI channel.

Observation 27: D2R sequence design and 2SB signals could be used for coarse SFO estimation. 



Proposal 1: Support both of Alt 1 (Single ON-OFF transmission) and Alt 2(multiple ON-OFF transmission) for SIP pattern design.

Proposal 2: Support Alt 1-1 and Alt 1-3 of single ON-OFF transmission for SIP design.
Alt 1-1: ON followed by OFF with same duration for both
Alt 1-3: ON followed by OFF with a duration ratio of [2,3]:1

Proposal 3: Support Alt 2-1 of multiple transmission for SIP design.

Proposal 4: Do not support Alt 2-3 (OFF-ON-OFF) of multiple ON-OFF transmission for SIP design.

Proposal 5: Support Alt 2-2 (ON-OFF-ON) of multiple ON-OFF transmission for SIP design.

Proposal 6: Support the entire length of SIP set at most one NR symbol length for device 1.

Proposal 7: Consider the impact of transition edge period on CP detection of CAP for edge/transition detection at device side.

Proposal 8: Down select 1.5 OFDM symbol duration as CAP entire length for transmission.

Proposal 9: Discuss low-rate CAP or high-rate CAP for different coverage and access requirements for option 1

Proposal 10: It is preferred that the number of ON/OFF transmissions in the CAP is not fixed.

Proposal 11: Except for ON-OFF or ON-OFF pattern to meet the requirement of option 1or option 2, like SIP pattern could be considered for CAP pattern design, including ON-OFF-ON and OFF-ON-OFF, etc.

Proposal 12: If CAP occupy multiple OFDM symbol, consider how/whether to design CP insertion for subsequent OFDM symbol.

Proposal 13: For calculating R2D chip duration based on CAP transition edge, 
Support CP inserted used for chip duration calculation if the level of CP is different from that of following pattern in CAP.
Support at least following methods to calculated R2D chip duration
For ON-OFF/OFF-ON repetition (e.g., two time repetition), R2D chip duration=T1/2;
For ON-OFF-OFF-ON/OFF-ON-ON-OFF, R2D chip duration=T1/3 or T1/4, which up to next one chip.

Proposal 14: Do not support implicitly indication or violation of Manchester coding rule for ensuring end of PDRCH transmission.

Proposal 15: For D2R preamble pattern and sequence design, consider following table as baseline for specification study.
 

Proposal 16: Consider other OOK sequences without line code to meet the requirement of channel/interference/SFO estimation at reader side, e.g., [111..11] or [000111...] or other types.

Proposal 17: Consider other BPSK sequences without line code to meet the requirement of channel/interference/SFO estimation at reader side, e.g., [+1,+1,+1..+1,+1] or [+1,+1,+1,-1,-1,-1...] or other types.

Proposal 18: Support M sequence and consider maximum sequence length/initial sequence for M sequence.

Proposal 19: When Msg 1/Msg 3 transmits data, D2R Midamble can be used for synchronization tracking or SFO estimation. D2R frame structure could consider as below
Option 2: D2R preamble + X midamble(s), where X≥1
Option 4: D2R preamble + Y midamble(s) + postamble, where Y≥1
Note: FFS X and Y values based on TBS and the number of Msg 1/3 during one D2R signal.

Proposal 20: D2R Midamble is not necessary when device 1 only transmits energy state with 1bit information because of the small D2R message. Thus, Option 1 or Option 3 could be as potential D2R frame structure.

Proposal 21: Partially transmitting Preamble/Midamble/Postamble during one Msg 1/3 set for one inventory/command service to estimate channel or interference or SFO could be considered to reduce transmission overhead at device side and detection overhead at reader side.

Proposal 22: Consider sequence design to help reader to estimate CLI channel more accurately, including
Quiet period setting by successive OFF chip or bit ‘0’ modulation without line coding
Sequence design using BPSK or OOK with line coding
FFS how to design binary sequence (e.g., +1/-1) to help reader to estimate CLI channel

Proposal 23: For D2R midamble, whether supporting D2R Midamble or not, it is up to Msg 1/Msg 3 types (e.g., w/o data carrying) and TBS.

Proposal 24: Limited by device’s TBS, D2R postamble should be at least deprioritied for device 1 for the indicator of end of PDRCH and other functionality.

Proposal 25: Consider D2R Midamble be as the indication of the end of PDRCH transmission, at least including special sequence design and the gap consideration between last D2R Midamble and previous one.
Proposal 26: For D2R transmission, consider the potential SFO estimation methods at reader side by carefully design D2R sequence and 2SB signal.

3GPP TSG RAN WG1 #120bis													R1-2501762
Wuhan, China, April 7th – 11th, 2025
Title: 	Discussion on Ambient IoT signals
Source: 	ZTE Corporation, Sanechips
Agenda item:	9.4.3
Document for:	Discussion and decision

Conclusion
In this contribution, the downlink and uplink signal/channels are discussed for A-IoT. We have the following observations and proposals.
Observation 1: If a large M value is used in SIP, the required PRB number will be large.
Observation 2: In RFID, the duration of delimiter is 12.5us, which is slightly greater than a chip length of M=6.
Observation 3: A large M value is more sensitive to the CP handling method and may impact the detection of SIP, thereby affecting performance.
Observation 4: SIP which does not occupy a whole number of OFDM symbols (for instance, occupying 1/2 of an OFDM symbol) is not applicable for CAP/data with M=1.
Observation 5: For M=1 and M=2, the first OFDM symbol will contain more ON transmission than subsequent OFDM symbols, leading to a power imbalance issue.
Observation 6: For SIP, ‘1 1 0 0 0 0’ with M=6 and ‘1 1 1 1 0 0’ with M=6 perform better MDR and FAR.
Observation 7: With option 2, where ‘the duration of the clock-acquisition part remains consistent across all M values’, at least three OFDM symbols are required for clock-acquisition part. This option would be excessive and not resource-efficient for large M.
Observation 8: 3 chips in CAP is sufficient for small M values. However, for large M values, it is inadequate.
Observation 9: At the same SNR, as the number of chips in CAP increases, the FDR decreases.
Observation 10: 3 chips in CAP is not sufficient for M=12 and M=24.
Observation 11: With same number of chips in CAP, the required SNR for a target BLER of 10% and 1% are different for different M values.
Observation 12: If the M value of the control information differs from that of the R2D data, R2D chip may across the boundary of OFDM symbol.
Observation 13: When the same amble(s) overhead is maintained, adding a midamble provides comparable performance results to only using a preamble when TBS is 16.
Observation 14: When midamble(s) is used, the more bits left for the last part, the worse the performance.
Observation 15: Adding a midamble can provides at least 4dB performance gain at BLER = 0.1, compared to using only a preamble when the TBS is 96 bits
Observation 16: At least one midamble can be considered for TBS is 96 bits.
Observation 17: At least a preamble and 2 midambles should be supported in D2R transmission for TBS=400bits.
Observation 18: At least a preamble and 3 midambles should be supported in D2R transmission for TBS=1000bits.
Observation 19: Using an 8-bit or 16-bit sequence to generate a long sequence(e.g., 32 bits) by repetition can obtain similar performance with a long sequence.

Proposal 1: The M value for SIP should not be greater than 6.
Proposal 2: The SIP duration should be 1 OFDM symbol.
Proposal 3: For SIP, ‘1 1 0 0 0 0’ with M=6 or ‘1 1 1 1 0 0’ with M=6 is preferred.
Proposal 4: When designing the clock-acquisition part, the following factors should be taken into account:
- the length of the clock-acquisition part,
- the CP handling.
Proposal 5: Option 1 ‘Duration of the clock-acquisition part is variable for different M values, i.e. the duration becomes shorter with increasing value of M’ is preferred.
Proposal 6: CAP with a pattern of multiple alternating high and low voltages(i.e., 101010... or 01010...), the choice between using 101010... or 010101... is dictated by the voltage of the last chip in start indicator part.
Proposal 7: CAP with an odd number of chips , with a minimum of three chips, can be a start point.
Proposal 8: For CAP design, at least two different chip numbers in CAP are needed to ensure detection accuracy of M values.
- E.g., 3 chips of CAP for M less than 12, and 5 chips of CAP for M greater than or equal to 12.
Proposal 9: Using same M value for both control information and R2D data, if control information is supported in PRDCH.
Proposal 10: For R2D postamble pattern, the following two options can be considered:
- Option#1: high voltage
- Option#2: low voltage
- Note : The length of R2D postamble should be greater than two chips.
Proposal 11: The voltage of R2D postamble can be same as the padding bits used for the last OFDM symbol.
Proposal 12: Multiple sequence lengths can be studied for the D2R preamble/midamble.
Proposal 13: In one D2R transmission, the length of preamble and midamble can be same.
Proposal 14: ' Preamble only' can be considered at least for a small TBS, e.g., less than 16 bits or 20 bits.
- A sequence with 32/64 bits for D2R preamble can be a start point when only preamble is used.
Proposal 15: The number of midamble(s) should be indicated by R2D control information, and device determine the position of midamble(s) according to the number of midamble(s) and predefined rule.
Proposal 16: The following predefined methods can be considered to determined the position of midamble(s):
- Alt#1: Insert one midamble at the end of PDRCH, and uniformly insert other midambles in addition to the preamble and the last midamble.
- Alt#2: Insert one midamble at the Pth-to-last bit position, and uniformly insert other midambles between the preamble and the last midamble.
Proposal 17: For D2R midamble, the last midamble should be close to the end of the transmission.

Proposal 18: For D2R midamble, at least 1,2,3 midamble(s) should be supported for D2R transmission.
- If repetition is employed for D2R transmission, more midamble(s) should be considered.
Proposal 19: Consider the following options for X-amble sequence generation:
- Option1：For M sequence, generating long sequences by repeating short sequences.
-Option 2：For Golay sequence, generating long sequences according to a pair of Golay complementary sequences or repeating.

3GPP TSG RAN WG1 #120b	R1-2501778
Wuhan, China, April 7th - 11th, 2025

Source:	Nokia
Title:	AIoT Timing acquisition and synchronization
Agenda item:	9.4.3
Document for:	Discussion and Decision

Conclusion

In this contribution we have made the following observations and proposals:
List of Observations
	Start delimiter should be long enough to ensure reliable detection due to large sampling offset observed in the AIoT tag before estimating/correcting from preamble.
Observation 1:
Observation 2:	SIP design should consider level detection followed by edge detection to ensure reliability while expending lower power.
Observation 3:	Preamble sequences must have better autocorrelation properties to obtain better timing accuracy even with large sampling offset.
Observation 4:	Increasing the value of M limits the coverage of transmission but improves the timing accuracy.
Observation 5:	Increasing the value of M, while retaining the same transmission duration, improves the auto-/cross- correlation properties of the sequence due to increase in sequence length.
Observation 6:	Having a shorter CAP sequence may trigger higher FAR for higher M values than the one with lower M values.
Observation 7:	RAN1 to consider designing PRDCH preamble by ensuring uniform distribution of 1’s and 0’s, i.e., by simply repeating 1,0 pattern to the desired length.
Observation 8:	Use of  provides better timing estimation accuracy even in the presence of SFO as the sequence length spans only over a shorter duration.
Observation 9:	The length of preamble sequence must be decided considering coverage, timing estimation accuracy, and the robustness against SFO.
Observation 10:	As the Manchester encoding is assumed for PRDCH payload, midamble may not be needed as Manchester coding itself is a self-synchronizing sequence.
Observation 11:	The use of midamble is needed to reacquire the timing synchronization for the device having large sampling offset, if the duration of the transmission lasts longer following the preamble.
Observation 12:	Even with large sampling offset, midamble may not be needed if the payload length is short, since the initial timing alignment is carried out by the preamble sequence.
Observation 13:	Terminating the PRDCH transmission with a postamble may provide two benefits, namely, the variable payload length and to provide timing acquisition before the subsequent transmission of either PDRCH or PRDCH, thus improving the detectability at both reader and the device, respectively.
Observation 14:	Using postamble and preamble together with the length of PRDCH transmission, an AIoT device can determine the sampling offset/drift experienced.
Observation 15:	The choice of preamble sequence, i.e., m-sequence or Golay sequence, may not affect the correlation performance as both are formed from primitive polynomial sequence family.
Observation 16:	The length of the preamble for PDRCH preamble can be limited to 32 or 48 as the reader can utilize coherent detection to improve the detectability.
Observation 17:	The need for midamble depends on the type of modulation used for D2R transmission. If the D2R transmission uses self-synchronizing modulation schemes like FM0, Miller, or MC, then the need for midamble is not essential.
Observation 18:	Since the residual SFO after receiving PDRCH preamble at the reader is less than 500ppm, the use of midamble may not provide significant detection improvement if Manchester encoding is used for data.
Observation 19:	As FEC is adopted in D2R transmission, device cannot autonomously terminate the transmission by sending postamble for early termination, since the coding rate is fixed by the size of the information bits and the codeword length, both are signaled in the previous PRDCH control.
Observation 20:	The postamble may improve detection if it is used as stop bits to ensure reception of all payload bits in PDRCH before FEC decoding when there is larger SFO at the AIoT device.

List of Proposals
Proposal 1:	RAN1 shall consider SIP using Alt-2 shall be favored, in particular Alt-2-1.
Proposal 2:	RAN1 to consider designing PRDCH preamble using alternate 1’s and 0’s.
Proposal 3:	RAN1 should consider a preamble sequence of length at least 32 to ensure reliable detection and to lower FAR.
Proposal 4:	Consider fixed transmission duration irrespective of the value of M, i.e., option 2 is preferrable, i.e., the sequence length will scale by the value of M.
Proposal 5:	As Manchester encoding is considered for R2D transmission, the need for midamble is not imperative.
Proposal 6:	RAN1 shall not consider postamble for PRDCH transmission as the reader is controlling both transmission size and power level.
Proposal 7:	RAN1 should consider generating preamble sequence for D2R transmission using m-sequence as it is simpler to implement at AIoT device and has better auto/cross correlation properties.
Proposal 8:	RAN1 should consider D2R preamble sequence of length of 16- or 32-bits using m-sequence.
Proposal 9:	Due to Manchester encoding like ON/OFF pattern used for PDRCH, RAN1 shall refrain from using midamble.
Proposal 10:	The use of postamble for PDRCH transmission shall be signaled by the reader depending on the payload size and the SFO accuracy of the AIoT device.

3GPP TSG RAN WG1 #120 bis	R1-2501808
Wuhan, China, April 7th – 11st, 2025

Source:	vivo
Title:	Discussion on Timing acquisition and synchronization for AIoT
Agenda Item:	9.4.3
Document for:	Discussion and Decision

Conclusion
In this contribution, we provide our views on R2D and D2R amble design, and the observations and proposals are provided as follows.
R2D
Observation 1: The device can not measure the absolute time of predefined chip during SIP detection due to uncalibrated local clock in SIP detection.
The initial clock accuracy may worse than 10^6 ppm.
Observation 2: The reasonable way to detect SIP is to detect the relative length ratio of ON state and OFF state, e.g., to identify (more than) 3:1 pattern in R2D detection, once the target relative length ration is detected, the SIP is assumed as detected at AIoT device.
Observation 3: M>6 for R2D SIP should be avoided, to make sure AIoT R2D can be deployed within Tx BW of one RB.
Observation 4: Single ON/OFF transmission does not work if there is RF/CW signal preceding SIP transmission.
Observation 5: When M=12, the ratio of ON/OFF length is very close to 1:3, which makes it difficult to be distinguishable from SIP pattern.
Proposal 1: At least 3:1 of OFF/ON length ratio should be defined for SIP.
M=12 is excluded from M value set to avoid ambiguity with SIP pattern.
Proposal 2: M value is 6 for SIP part, and time duration for SIP is one OFDM symbol.
Observation 6: Alt2-2 achieves best MDR with lowest FAR rate.
Proposal 3: Support Multi-ON-OFF pattern with Alt2-2.
Duration for SIP signal is one OFDM symbol using M = 6 and 100001.
Proposal 4: For the CAP of the R2D, support Option 1 using 4 ON-OFF chips with ‘OFF/ON/OFF/ON’. 
The maximum duration for CAP is two OFDM symbols, corresponding minimum M value of 2.
Proposal 5: No additional signal design is needed to handle the CP between SIP part and CAP part.
Proposal 6: Based on RAN2 agreements, R2D postamble is not supported.
D2R
Observation 7: The amble options should be evaluated assuming fading channel for both CW2D link and D2R link. Otherwise, parameters for ambles, e.g., density and sequence length, may be underestimated if fading channel is considered only for D2R link.
Proposal 7: The sequence used for pre/mid-ambles are Manchester coded and frequency shifted in the same way as the corresponding data part.
Observation 8: Two parts sequence using all ones as first part can achieve better performance than one part sequence, with low complexity, i.e., less number SFO hypothesis in correlation.
Observation 9: For the same preamble length, processing complexity of preamble with two parts is up to ~84% lower compared with that of preamble with one part.
Observation 10: In existing RFID specs, i.e., 18000-61/62/63(UHF RFID C1Gen2)/64, preamble is composed by two parts, starting by all 0/1s which is followed by sync sequence.
Proposal 8: Preamble is composed by two parts.
Part 1 is all 1s or all 0s sequence (before Manchester encoding).
Part 2 is Golay or M-sequence.
Observation 11: For the same length of preamble, better performance can be achieved when preamble part 1 is not shorter than preamble part 2.
Proposal 9: Multiple preamble length can be supported and indicated by reader. 
Reader controls the preamble length based on reliability requirement.
Observation 12: Performance difference between M-sequence and Golay sequence is marginal, while Golay sequence has more choices in length, for length < 100 the following length can be used for Golay and M sequence, respectively
Golay sequence Length= {8,10,16,20,26,32,40,52,64,80};
M sequence Length= {7,15,31,63}.
Observation 13: For ~5kbps and ~10kbps data rate, the absolute time duration for D2R data part beyond which additional midamble is required to improve D2R performance at 1% BLER are similar, which is about 8 ms.
Proposal 10: Around L=44 bits (before 1/3 channel coding) with ~5kbps data rate can be considered as a reference payload size to determine whether additional midamble is needed
-  L can be scaled to apply to other data rate configurations.
Proposal 11: Depending on reliability requirements, it can be up to reader to determine whether additional midamble is transmitted even when length of D2R data part is longer than certain threshold.
Proposal 12: Parameters for midamble, e.g., length, number of midambles in D2R transmission, can be controlled by reader.
{midamble number N, length of midamble in number of bits L} depending on reliability requirements.
Proposal 13: The following methods can be considered to determine the location of midamble(s).
Alt-1: The last midamble location is indicated and the other midambles are uniformly distributed between preamble and last midamble.
Alt-2: The midamble location is determined based on gap between midambles except for the last midamble, and whether there is last midamble can be indicated separately.
Alt-3: Uniformly distribute midambles (except for the last midamble), and whether there is last midamble can be indicated separately.
Observation 14: Motivation of multiple sequences that Multiple AIoT devices use different sequences for concurrent D2R transmission to different readers is not valid.
Observation 15: It is not clear how randomization can provide any benefit if the two devices are already FDM multiplexed in different frequency shift.
Proposal 14: For a given sequence length, do not support multiple M/Golay sequences, single sequence is sufficient.

3GPP TSG RAN WG1#120bis                                                               R1- 2501870
Wuhan, China, April 7th – 11th, 2025

Agenda Item:     9.4.3
Source:	Spreadtrum, UNISOC
Title:	Discussion on timing acquisition and synchronization for Ambient IoT
Document for:	Discussion and decision

Conclusions
Obsevation1: For Alt1-2 that ON followed by OFF with a duration ratio of 1: 3, the MDR can achieve below 2%, and FAR can reach 0.2%. Therefore, the reliability of Alt 1-2 is enough, and the “OFF” duration can be distinguished from subsequent CAP and PRDCH.
Proposal 1: For start indicator part in the R2D preamble:
Alt 1-2 should supported that ON followed by OFF with a duration ratio of 1:3, and the length of start indicator part is fixed to one OFDM symbol for all M values of PRDCH.
Proposal 2: For clock-acquisition part in the R2D preamble, Option 1 is adopted with the following:
Duration of the clock-acquisition part is variable for different M values, i.e. the duration becomes shorter with increasing value of M.
The maximum duration of CAP is 3 OFDM symbol duration.
The CAP is to indicate the chip length for the both data and control information transmitted in the subsequent PRDCH. 
Proposal 3: CP handling Alt M2-1-1 can be used for M=4, 16, 24.
When M = 4, 16, 24, the fixed pattern of CAP should be ON-OFF-ON-OFF, the two OFF parts is used to determine the chip duration. 
Proposal 4: To determine or derive the end of PRDCH transmission:
For small payload size, TBS information is explicit indicated by L1 R2D control information.
For large payload size, postamble is used to indicate the end of PRDCH.
Proposal 5: Postamble is supported for R2D transmission, and the sequence pattern is different from preamble and PRDCH.
More than 3 consecutive high-voltages pattern is used for postamble.
Proposal 6: For the sequence design of D2R preamble, autocorrelation needs to be considered, and multiple orthogonal sequences are not supported.
Proposal 7: Short sequence and long sequence of D2R preamble should be supported for different D2R payload size.
Proposal 8: For D2R midamble, the sequence of Midamble is the same as preamble.
Short sequence and long sequence should be supported.
The number of Midambles is dynamically indicated by the reader in the PRDCH.
Proposal 9: If the number of midamble is one, midamble is transmitted at the end of PDRCH by default. If the number of midamble is larger than one, one is transmitted at the end of PDRCH, the others are transmitted during the PDRCH transmission controlled by the Reader.

3GPP TSG RAN WG1 #120bis                                                                                 R1-2501895
Wuhan, China, April 7th – 11th, 2025

Agenda Item:	9.4.3
Source:	Lenovo
Title:	Discussion on timing acquisition and synchronization for Ambient IoT
Document for:	Discussion 

Conclusion
In this contribution, we focus on  timing acquisition and synchronization for ambient IoT and have observations and proposals on preamble/midamble/postamble for R2D and D2R, respectively as follows:

Proposal 1: For the SIP of R-TAS, prefer to support Alt1-2 or Alt1-3 from the device side. However reader can be configured to transmit single ON-OFF or multi ON-OFF transmission using Alt 2-1. 
Proposal 2: Support duration of the clock-acquisition pat is constant for different M values (Option 2) for clock-acquisition part of R2D transmission. 
Proposal 3: Consider clock acquisition part design containing different M values with repetition such that larger M value tolerates large timing errors at the beginning to receive robust L1 control followed by smaller M value for finer synchronization needed for the R2D data reception.
Proposal 4: Compare the overhead of TBS indication vs. postamble size to decide on the method indicating the end of PRDCH transmission.
Proposal 5: Support binary Golay sequence to be used for D2R preamble.
Proposal 6: Support D2R preamble carrying information to facilitate decoding of repetition based D2R transmission.
Proposal 7: A criteria (e.g., length threshold of D2R transmission) can be defined for determining whether/how to place midamble in D2R transmission.
Proposal 8: Midamble related information (e.g., length threshold of D2R transmission) can be indicated in one of the following options:
Option 1. by the corresponding R2D in the D2R scheduling information.
Option 2. in a pre-defined way.
Proposal 9: On the length of D2R midamble, one of the following alternatives can be considered. 
Alt 1. One single length midamble
Alt 2. Multiple lengths midambles
Proposal 10. Support using D2R midamble resource for interference estimation. 
Proposal 11. Whether D2R midamble resource is reused for interference estimation needs to be indicated to device by reader. 
Proposal 12: For the repetition-based D2R transmission, support to acquire the end of transmission of each repetition by using midamble following the repetition.

3GPP TSG RAN WG1 #120bis                     	R1-2501921
Wuhan, CN, April 7th – 11th, 2025

Agenda Item:	9.4.3
Source:	InterDigital, Inc.
Title:	Discussion on timing acquisition and synchronization
Document for:	Discussion and Decision

Conclusion

The following observations are provided for discussion in RAN1#120bis:

Observation 1:	There is no clear trend with regard to preamble design dependent on small-frequency shift used.
Observation 2:	M-sequences and Golay sequences offer comparable performance with regard PDRCH BLER, and SFO estimation.
The following proposals are provided for discussion in RAN1#120bis:

Proposal 1:	For the CAP of R-TAS, the values of M supported are the same as the values of M supported for the PRDCH transmission.
Proposal 2:	The M value used for the CAP of R-TAS is the same as the M value used for the PRDCH transmission which follows it.
Proposal 3:	For the CAP of R-TAS, support option 1, the duration of the CAP changes with changing M used for the PRDCH.
Proposal 4:	D2R preamble is based on m-sequence design similar to PSS from Uu interface.
Proposal 5:	D2R preamble does not include a clock pattern portion as part of its design.
Proposal 6: 	D2R preamble and midamble are configured separately by the reader. 
Proposal 7: 	At least two formats are supported for D2R preamble and midamble.
Proposal 8: 	Format and time domain resources for D2R midamble are indicated as L1 control by the reader.

3GPP TSG RAN WG1 #120bis                                                                            R1-2501994
Wuhan, China, April 7th – 11th, 2025

Source:             CATT
Title:             	Ambient IoT timing and synchronization
Agenda Item:    9.4.3
Document for:  Discussion and Decision

Conclusions
In this contribution, we discuss timing acquisition and synchronization design in support of Ambient IoT devices, and give the following observations and proposals:
Observation 1: The ON/OFF pattern of SIP should be designed considering both scenarios that the energy harvest for A-IoT charging or other R2D transmission may or may not take place immediately before the SIP.
Observation 2: The device needs to detect at least one OFDM symbol of the SIP, in order to meet the SIP detection requirement of target MDR of up to 1% for FAR of up to 1%. 
Observation 3: The SER will be degraded a lot due to un-even chip interval when the CP is inserted within an OFDM symbol.  
Observation 4: The device needs to detect at least 6 or 4 chips in the CAP, in order to obtain about 15% or 20% SER performance, respectively. SER refers to the number of samples which is mismatched for comparing to the total number of samples in a chip.
Observation 5: For the duration of the CAP for different M values:
The synchronization performance would be the same for all M-values of OOK-4 waveform when the duration of the CAP is variable with the number of “0” and “1” in the CAP being fixed for different M values, i.e. the duration becomes shorter with increasing value of M.  The residual timing errors would degrade the OOK-4 demodulation performance more with large M-value when the same number of bits “0” and “1” in the CAP sequence
The synchronization performance would improve in proportion to the increase of the M value of OOK-4 waveform when the CAP duration is fixed and the number of the “0” and “1” in the CAP sequence increases in portion to the increase of the M-value. The OOK-4 demodulation performance would be similar for different M values as the impact of the residual synchronization errors decrease to the performance of OOK-4 demodulation when the chip duration decreases in proportion to the increase of the large M value.  
Observation 6: For the duration of the CAP for different M values, down-select from the following options: 
Option 1-1: If duration of the CAP is variable for different M values and M=1 is supported, the CAP includes 3 chips for all candidate values of M (i.e., 1, 2, 4, 8, 16). The sequence in the CAP for all candidate values of M is fixed to {0, 1, 0}. 
Option 1-2: If duration of the CAP is variable for different M values and M=1 is not supported, the CAP includes 4 chips for all candidate values of M (i.e., 2, 4, 8, 16). The sequence in the CAP for all candidate values of M is fixed to {0, 1, 0, 1}.
Option 2-1: If duration of the CAP is constant for different M values based on repetition and M=1 is supported, the duration of CAP is fixed to 3 OFDM symbols with the {0, 1, 0, 1…} sequence for all the candidate values of M.
Option 2-2: If duration of the CAP is constant for different M values based on repetition and M=1 is not supported, the duration of CAP is fixed to 2 OFDM symbols with the {0, 1, 0, 1…} sequence for all the candidate values of M.
Observation 7: The counting of new transition edge of CP would degrade the synchronization performance and increase the A-IoT device complexity.
Observation 8: For the cases of D2R preambles:
The performance difference for sequence type of D2R preamble with M sequence and Golay sequence is very small at the BLER point of 10-1 and 10-2.
Compared to Part 2 sequence length of preamble sequence, Part 1 sequence length of preamble sequence has a greater impact on BLER performance, mainly due to part 1 sequence length of preamble sequence impact on SFO estimation performance.
Observation 9: For the case of the payload size being 96 bits,
D2R midamble can marginally improve the performance of SFO estimation, timing acquisition and D2R transmission BLER slightly.
The performance with one or two D2R midambles is almost same at the BLER point of 10-1 and 10-2.
Observation 10: For the case of the payload size being 400 bits,
D2R midamble can improve the performance of D2R transmission performance.
The performance with one or two D2R midambles is almost same at the BLER point of 10-1 and be less than 1dB at the BLER point of 10-2. 
Proposal 1: For A-IoT devices, the R2D transmission or carrier wave may or may not take place immediately before R-TAS transmission.
Proposal 2: The guard period can be added at the beginning of the SIP ON/OFF pattern in order to avoid the impact of energy harvest for A-IoT charging may or may not take place immediately before the SIP.
Proposal 3: The SIP of R-TAS should be multi-ON-OFF transmission with ON-OFF-ON or OFF-ON-OFF (duration of ON and OFF can be different). The pattern of the SIP should be different from that of subsequent clock-acquisition part of R-TAS and PRDCH transmission.
Proposal 4: The SIP of R-TAS should include one OFDM symbol to meet the SIP detection requirement of the target MDR of up to 1% for FAR of up to 1%.
Proposal 5:  The A-IoT synchronization performance by CAP of R-TAS would be degraded with SER if CP is included as part of OOK-4 waveform. SER refers to the number of samples which is mismatched for comparing to the total number of samples in a chip.
Proposal 6: Only the values of M = {2, 4, 8, 16} should be supported for the CAP of R-TAS in Rel-19.
Proposal 7: For the duration of the CAP for different M values, Option 2-2 should be supported: 
Option 2-2: the duration of the CAP is constant for different M values based on repetition and M=1 is not supported, the duration of CAP is fixed to 2 OFDM symbols with the {0,1,0,1…} sequence for all the candidate values of M.
Proposal 8: The chip duration of CAP of R-TAS and the chip duration of PRDCH should be the same, and the reader only needs to indicate a unique R2D chip duration for both CAP of R-TAS and PRDCH.
Proposal 9: For the design of R-TAS, there is no need to introduce any other part for R-TAS except for the SIP and CAP.
Proposal 10: Method Type 2-M2-2, which extends the legacy OFDM symbol duration without NR CP operation, should be supported for R2D signal transmission due to its lower complexity and more efficient spectrum resource utilization compared to other methods.
Proposal 11: If the orthogonal method is adopted, a candidate method in Method Type 2 (Alt M2-1-1) could be supported, i.e. the end of an OFDM symbol corresponding to the CP length will be set to high voltage transmission for M equals to 16.
Proposal 12: R2D postamble should not be introduced to indicate the end of PRDCH transmission. TBS information should be used to implicitly indicate the packet size and transmission time interval and could be included in L1 R2D control information.
Proposal 13:  Joint CRC should be applied to both R2D control information and R2D data.
Proposal 14: TBS indication, which is not included in the CRC calculation, should be protected by simple error correction and error detection mechanism of simple sequence detection, such as Walsh sequences.
Proposal 15: The reader can obtain D2R chip synchronization by detecting the rising and falling edges of the D2R CAP sequence.
Proposal 16: D2R preamble can support multiple sequences for collision avoidance in future release.
Proposal 17:  The D2R preamble design needs to take into consideration of performing fine frequency synchronization and clock calibration for Device 2b which may be supported in future release.
E.g., a single-tone sine wave can be transmitted before the D2R preamble for Device 2b.
Proposal 18: For A-IoT devices, at least R2D signal can be used to assist A-IoT devices in adjusting the transmission timing of PDRCH to battle the initial SFO from local oscillator and reducing the residual SFO of A-IoT devices.
Proposal 19: From the perspective of robustness of TBCC to SFO, D2R preamble design should ensure that residual SFO does not exceed 1000ppm.
Proposal 20: The D2R preamble with part 2 sequence can be either an M-sequence or a Golay sequence.
Proposal 21: Part 1 sequence length of D2R preamble sequence should not be less than 64 chips, and the part 2 sequence length of D2R preamble sequence can be less than 64 chips.
Proposal 22: The chip duration of D2R preamble and the chip duration of PDRCH should be the same, and the reader only needs to indicate a unique D2R chip duration for both D2R preamble and PDRCH.
Proposal 23: For different chip durations of D2R preamble, the length of D2R preamble should be the same and the number of chips of D2R preamble should be different.
Proposal 24: If D2R TB size is larger than a pre-defined condition, such as the threshold value of the TBS size, D2R midamble is present; else D2R midamble is not present.
Proposal 25: All the transmission parameters of D2R midamble should be pre-defined.
Proposal 26: At least the following parameters of D2R midamble should be specified:
The position of the first D2R midamble during the PDRCH transmission.
The duration of D2R midamble.
The interval between two adjacent D2R midambles.
Proposal 27: Similar to D2R preamble sequence design target, D2R midamble sequence design should ensure that residual SFO does not exceed 1000ppm and should improve the BLER performance @ 10-1 and 10-2 obviously.
Proposal 28: For D2R transmission with larger payloads (e.g., 400 bits), the performance of D2R transmission can be improved through the insertion of D2R midamble.
Proposal 29: For the reader to acquire the end of PDRCH transmission, the Option 2 should be adopted:
Option 2: Based on the corresponding R2D control information.

3GPP TSG RAN WG1 #120bis			R1-2502024
Wuhan, China, April 7th – 11th, 2025

Agenda item:		9.4.3
Source:	China Telecom
Title:	Discussion on timing acquisition and synchronization for Ambient IoT
Document for:		Discussion

Conclusions
In this contribution, we have the following proposals:
Proposal 1: For start-indicator part design, we support Alt 1-1, i.e., ON followed by OFF with same duration for both.
Proposal 2: Support using 1 OFDM symbol duration for the entire duration of SIP, i.e., 1/2 OFDM symbol duration for ON or OFF transmission.
Proposal 3: Support 3 OFDM symbol duration for maximum duration of CAP.
Proposal 4: Support option 1 that variable CAP durations for different M values while maintaining a fixed number of ON/OFF transmissions. 
Proposal 5: Support indicating the end of PRDCH transmission via L1 R2D control information and R2D postamble. 
When the PRDCH transmission duration is less than X, indication via L1 R2D control information is recommended. Conversely, for transmission durations exceeding X, indication via postamble is a viable alternative. FFS: the value of X.
Proposal 6: Design postamble based on violation of Manchester coding rule, and “ON-ON-ON-OFF” can be used as a baseline for postamble design.
Proposal 7: Support using multiple sequences with the same base sequence in D2R x-ambles.
Proposal 8: For D2R transmission, preamble-only option is supported (attached before the start of PDRCH), when the PDRCH payload size is up to X bits
-FFS the value of X
Note: For determining value of X, companies are encouraged to evaluate the designs in terms of SFO estimation, and corresponding PDRCH BLER
Proposal 9: Distinguish the design between preamble and midamble.
Proposal 10: The presence and detailed pattern of midambles for D2R transmission can be indicated by R2D control information.

3GPP TSG RAN WG1 #120-bis	R1-2502042
Wuhan, China, Apr. 7th - 11th, 2025

Source:	Ofinno
Title:	Views on Timing acquisition and synchronization
Agenda item:	9.4.3
Document for:	Discussion and Decision

Conclusions
In this paper we make the following observations and proposals: 
 Proposal 1: For the SIP of R-TAS, support:
Alt 1. Single ON-OFF transmission 
Support one of the following
Alt 1-1: ON followed by OFF with same duration for both
Alt 1-3: ON followed by OFF with a duration ratio of [2,3]:1
Proposal 2: Support 1/2 OFDM symbol duration for the duration of the SIP.   
Observation 1: CP handling is needed for the proper reception of the CAP and determination of M value. 
Proposal 3: RAN1 to wait for more progress on CP handling before moving forward with the CAP design.
Proposal 4: RAN1 to wait on deciding the maximum length of the CAP until at least the minimum value of M is agreed. 
Proposal 5: Don’t support an additional part for the D2R preamble beyond the base sequence (e.g., M-Sequence or Golay sequence).    
Proposal 6: For the D2R x-able design, support multiple sequences (for the same base sequence) and support multiple sequence lengths. 

3GPP TSG RAN WG1 #120bis	R1-2502125
Wuhan, China, April 7th – 11th, 2025

Agenda Item:	9.4.3
Source:	Fujitsu
Title:	Discussion on timing acquisition and synchronization
Document for:	Discussion and decision.

Conclusion
In this contribution, we consider the aspects of timing acquisition and synchronization. We have the following observations and proposals:
Observation 1: For multi-ON-OFF transmission in SIP,
No need to utilize multi-ON-OFF transmission for calibrating the threshold for judgement of high voltage and low voltage.
The high-low voltage transmission in R2D preamble can be utilized to achieve this calibration goal.
The design of multi-ON-OFF transmission will cost more overhead and require more standardization efforts than single-ON-OFF transmission.
Proposal 1: Support single ON-OFF transmission (Alt.1 in Option 1) for SIP, which is enough for devices to recognize the start of R2D transmission and can save standardization effort.
Proposal 2: For single ON-OFF transmission of SIP, 
Predefine a minimum length for the ON duration in SIP, and
The OFF duration can be in a fixed length or a variable length within a range.
Proposal 3: For SIP of the R2D time acquisition signal,
To guarantee that devices can distinguish between the OFF duration of SIP and the lengths of OOK modulation chip with all M values ,
The length of OFF duration should be obviously different from one or two times the chip lengths with all M values.
Proposal 4: The length of SIP does not vary with the OOK M value of the following PRDCH. The length of ON-OFF duration in SIP is predefined.
Observation 2: For the duration of CAP for different M values, in the case that the pattern of CAP in Option 1 is assumed to be same for all values of M,
Option 1 may be more efficient than Option 2 and can save more resources and standardization efforts.
Option 2 can provide more accurate clock calibration than Option 1.
Proposal 5: Support Option 2, i.e., the duration of CAP is constant for different M values. 
Observation 3: For the pattern of CAP with constant duration for different M values, to include at least two transitions from “OFF” chip to “ON” chip or two transitions from “ON” chip to “OFF” chip,
Option 1, ON-OFF repetition pattern occupies more OFDM symbols resources than Option 2.
Option 2, ON-OFF-ON repetition pattern includes less transitions for clock tracking than Option 1. 
Proposal 6: If CAP with constant duration for different M values is supported, Option 1, i.e., ON-OFF repetition pattern is preferred. Option 1 provides more precision than ON-OFF-ON pattern.
Proposal 7: The M value for CAP is the same as the M value of its following PRDCH, the specific values of M follow the conclusion to be made in agenda 9.4.1.
No need to restrict M values for CAP to some certain values. 
Observation 4: The following issues occur when M value is large and the length of R2D chip gets comparable with the length of CP, 
It is difficult for devices to distinguish between a CP and normal OOK chips and determine the length and location of CP due to ultra-low capability.
The calibration of R2D chip duration in clock-acquisition part could be inaccurate.
Proposal 8: For the impact of CP insert on CAP, wait for the discussion of CP handling in agenda 9.4.1.
Observation 5: To determine the end timing of R2D transmission,
For Option 1 in TR 38.769, the end timing of R2D transmission can be determined based on TBS information, which is informed via implicit/explicit L1 R2D control information. This scheme can save signaling overhead and increase transmission efficiency.
For Option 2 in TR 38.769, the end timing of R2D transmission can be determined based on postamble of PRDCH. This scheme is more robust and flexible than the scheme of Option 1.
Proposal 9: Support Option 1 in TR 38.769 to determine the end timing of R2D transmission.
Observation 6: For the D2R midamble sequence design, the auto-correlation property of candidates is one of the key factors for down-selection.
Proposal 10: For the number of D2R midambles, the following schemes can be considered.
Option 1: the number of D2R midamble is limited to one for a certain payload size of PDRCH.
Option 2: multiple values of the number of D2R midamble are supported for a certain payload size of PDRCH.
Proposal 11: The location of the D2R midambles inserted into the PDRCH can be determined by the following options.
Option 1: The block size between two adjacent ambles in PRDCH is fixed as Nblock, the first midamble is inserted at the position of Nblock after the end of preamble, the second midamble is inserted at the position of Nblock after the end of the first midamble, and so on.
Option 2: The location of midamble is determined according to the number of midamble M, the position of the first midamble is , where  is the payload size of the PDRCH,  is the block size between two ambles.

3GPP TSG RAN WG1 #120bis			R1-2502162
Wuhan, China, April 7th – 11th, 2025 

Agenda item:	9.4.3
Source: 	CMCC
Title: 	Discussion on timing acquisition and synchronization
Document for:	Discussion and Decision

Conclusions
In this contribution, we provide our views on downlink and uplink channel/signal, and the following observations and proposals are made:
Observation 1: A-IoT has a much larger coverage and lower receiver sensitivity, the output voltage of the envelop detector of a high voltage transmission may not have significant distinctions from a low voltage transmission for a coverage-edge device.
Observation 2: In case of different coverage levels, it is important that SIP can include a dynamic threshold training part for a device to set an appropriate threshold for the comparator.
Observation 3: A high sensitivity and low power consumption reference threshold detector for the comparator can be applicable for A-IoT device 1. 
Observation 4: The proposed SIP pattern can achieve a target MDR and FAR <= 1% when CNR >= 18 dB.
Observation 5: The proposed CAP pattern can achieve a target MDR <= 1% when CNR >= 18 dB.
Observation 6: For CAP pattern ON-OFF-ON, when M = 2, without handling the CP insertion impact on timing acquisition, a performance degradation of 1 dB is observed.
Observation 7: For payload size of 20 bits, when compared to the no midamble case, using preamble + midamble for fine SFO estimation, where midamble locates near the end or at the end of PDRCH, significantly improves the link performance. At 10% BLER, assuming 7 kcps, 3 dB gain is achieved using preamble + midamble when compared to no midamble case.
Observation 8: Even for a payload size of 20 bits, preamble only is not sufficient, and midamble should be used for link performance improvements.
Observation 9: For 20-bit payload size, up to 40 hypothesis is enough for SFO determination. 
Observation 10: For 20-bit payload size, preamble only case is 3dB worse than using preamble + midamble (midamble is located at the end of PDRCH).
Observation 11: For 96-bit payload size and up to 40 hypotheses, using preamble only does not work for SFO =10% cases.
Observation 12: No significant difference in link performance is observed for midamble with different locations. Midamble locating at the end of PDRCH for SFO estimation provides slight gain over midamble locating near the end of PDRCH.
Observation 13: When compared to midamble locating at the end of PDRCH, the indication complexity and overhead is expected to be increased for midamble inserting near the end of PDRCH.
Observation 14: The link performance of Golay sequence and m-sequence does not differ much. M-sequence has a slightly higher merit factor and therefore provides a slightly better performance than Golay sequence.
Observation 15: For FDMA achieved by applying small frequency shift, the power of the 3rd and 5th harmonics is reduced by approximately 9.5 dB and 14 dB, respectively, compared to the main lobe. Considering the extended coverage in A-IoT, devices accessing in FDMA may suffer from severe near-far effect, and harmonic interferences may significantly degrade the FDMA performance.


Proposal 1: For SIP, Alt. 2 with multiple ON-OFF transmission is supported.
Proposal 2: The entire duration of SIP occupies 1/2 OFDM symbol and is fixed for all supported M values. The pattern of SIP is ON-OFF-ON-OFF-OFF-OFF.
Proposal 3: The OOK chip duration of the clock-acquisition part is the same as that of the subsequent PRDCH transmission.
Proposal 4: For the pattern of CAP, Option 1 is supported, i.e., the duration of CAP is variable for different M values and the duration becomes shorter with the increasing value of M.
Proposal 5: The following options can be considered for CAP:
Option 1: ON-OFF-ON. The maximum duration of CAP is 1.5 OFDM symbol. CAP is applicable to M > 1.
Option 2: ON-OFF-ON-OFF-OFF-ON. The maximum duration of CAP is 3 OFDM symbol. CAP is applicable to M = 2, 6, 24. 
Proposal 6: Use R2D postamble to indicate the end of PRDCH transmission should be the baseline.
Proposal 7: The R2D postamble pattern is a duration of high voltage longer than the continuous OOK ON chips in R2D time acquisition signal and PRDCH transmissions, e.g., ON-ON-ON.
Proposal 8: For D2R, midamble is present regardless of D2R payload size.
Proposal 9: If only 1 midamble is indicated, it is inserted at the end of PDRCH.
Proposal 10: If more than 1 midamble is indicated, 1 midamble is inserted at the end of PDRCH and other midambles are inserted evenly into PDRCH.
Proposal 11: For D2R ambles, the following agreement is also applied, where Tb corresponding to a bit of a D2R amble.
Agreement
For D2R transmission with small frequency shift +/- R/Tb Hz and R = Tb/(2 × D2R chip length):
the transmitted 2R chips for bit 1 and bit 0 are [0 1 0 1 …] and [1 0 1 0 …], respectively, for OOK modulation
the transmitted 2R chips for bit 1 and bit 0 are [-1 +1 -1 +1 …] and [+1 -1 +1 -1 …], respectively, for BPSK modulation.
Note: The case of R=1 is equivalent to using a Manchester line code without repeating each Manchester codeword.
Proposal 12: For D2R ambles, Golay sequence can be considered with the following property:
Low peak sidelobe level (PSL).
High merit factor (MF).
Zero cross-correlation zone (ZCCZ).
Proposal 13: For D2R ambles for timing tracking, a sequence length of 64 bits, i.e., 128 chips after Manchester encoding, can be considered as a starting point for low SNR cases, e.g., 10% BLER@-2dB .
Proposal 14: Optimizations on D2R amble design in consideration of harmonic interference for FDMA can be considered:
Option 1: D2R ambles of devices using adjacent frequency channels for access are TDMed.
Option 2: Golay complementary pairs or GCPs with large zero cross correlation zone can be used for D2R ambles of devices using adjacent frequency channels for access.

3GPP TSG RAN WG1 #120bis	R1-2502202
Wuhan, China, April 7th – 11th, 2025

Agenda Item:	9.4.3
Source:	NEC
Title:	Discussion on timing acquisition and synchronization
Document for:	Discussion and Decision
1	

Conclusion
In this contribution, we give our views on control and other aspects of ambient IoT. We propose that:
Proposal 1: Regarding ON/OFF pattern for SIP, only support option 1 in Rel-19.
Proposal 2: RAN1 to consider the following options for the start and end time of SIP:
Option 1: SIP starts from the start of an OFDM symbol;
Option 2: SIP ends at the end of an OFDM symbol;
Option 3: SIP may start at any time which satisfies the timing requirement, e.g., TD2R_min and TD2R_max.
Proposal 3: Regarding the duration of CAP for different M values, consider a compromise method that a CAP can have a first part with fixed duration for clock acquisition and a second part with variable duration for chip duration determination.
Proposal 4: for a PRDCH with relatively large M value (e.g., M>8), if CAP or part of CAP uses same M value as PRDCH, the CAP or part of the CAP with the same M value should avoid to cross OFDM symbol boundary.
Proposal 5: R2D postamble with [2, M] chips to meet the last OFDM symbol boundary.
Observation 1: It would be of benefit to support different D2R midamble sequence targeting different functionality. 
Observation 2: It would be of benefit to support different D2R midamble sequence targeting same functionality.
Proposal 6: Support multiple D2R midamble sequence design in Rel-19 AIoT.
Proposal 7: Time-domain and/or frequency-domain resource indication of D2R midamble can be considered.
Observation 3: Variable-length midamble design is more adaptive to actual PDRCH transmission.
Proposal 8: Support variable-length midamble sequence generated from fix-length midamble sequence in D2R midamble design. 
Proposal 9: When applying small frequency shift, the number of effective chip after small frequency shift should be multiples of R between nearby ambles.
Proposal 10: To avoid increase of amble overhead, the number of chip of amble should be multiples of R.
Proposal 11: Support omitting the second last D2R midamble when the gap between the second last midamble and the last midamble is smaller than a threshold.  
Proposal 12: Support midamble at the end of PDRCH when the remaining data is bigger than a threshold.

3GPP TSG-RAN WG1 Meeting #120bis                                                                        	  R1-2502235              
Wuhan, China, April 7- 11, 2025

Agenda Item:	9.4.3
Source:	Huawei, HiSilicon
Title:	Physical signals design
Document for:	Discussion and Decision

Conclusions
Based on the analysis in this contribution, we have the following observations and proposals:
Observation 1: A SIP with a fixed duration can avoid blind decoding and checking for SIPs of varying durations, and reduce the complexity of device implementation.
Observation 2: Half of an OFDM symbol is a suitable duration for the whole SIP, because the design of such length meets the performance targets and the remaining duration of the OFDM symbol can include an integer number of chips for any M candidate value.
Observation 3:  The duration of the ON-OFF transmissions is not needed to be equal to the chip length of PRDCH since its objective is for the device to easily distinguish the following signals from the SIP, such that it can definitively identify the start of the PRDCH transmission.
Observation 4:  There is no need to restrict the M value used for the SIP based on the corresponding minimum transmission bandwidth of 1 PRB since the device can adjust its LPF bandwidth to receive the CAP/PRDCH using other M values corresponding to different bandwidths.
Observation 5: The device is required to determine a detection threshold such that the device’s comparator can function properly.
Observation 6: If the device does not use a detection threshold determination part, it would have to use an inaccurate threshold based on previously received transmissions, which would result in a higher missed detection ratio and would impact the performance of the device.
Observation 7: A high-voltage part is necessary to be transmitted before the low-voltage part, in order for the device to detect a falling edge and determine the start of the low-voltage part.
Observation 8: The presence of other RF transmissions or noise prior to the transmission of the SIP does not impact the device’s ability to determine the detection threshold if a detection threshold determination part is used.
Observation 9: The design of Alt 1 does not have a provision for the device to determine the detection threshold, resulting in an increase in MDR. 
Observation 10: For Alt 2-2 and Alt 2-3, apart from the lack of a threshold determination part, the consideration of a guard period is unnecessary from our perspective and the duration of SIP based on these alternatives need to occupy a longer period than our design, which lowers the transmission efficiency.
Observation 11: If the duration of the CAP is fixed, the value of M used would have to be small, resulting in a longer fixed duration and would cause an increase in overhead and latency.
Observation 12: If the duration of the CAP is variable, it has the advantage of being shorter in duration with an increasing value of M.
Observation 13: If the duration of the CAP is variable, it would result in a higher transmission efficiency than when it is fixed, since a fixed duration would contain unnecessary edges especially when the value of M is large.
Observation 14: The FDR increases drastically after adding M=4, 8 to the M value set. 
Observation 15: The impact of CP to the detection threshold of the SIP is negligible for different CP lengths based on the simulation results.
Observation 16: The spectrum of the clock-like ‘[1, 0, 1, 0, … ]’ transmission is a sinc-based signal whose peak value is high and the peak value position varies with the chip length impact by SFO. Therefore, SFO can be estimated by finding the peak value position in the frequency domain.
Observation 17: The PDRCH performance with m sequence is similar to that with the Golay sequence.
Observation 18: Compared to the 32-bit or 64-bit preamble, the overhead of the 8-bit preamble is three to five times lower.
Observation 19: A fixed long length preamble would cause unnecessary overhead when the required SNR differs in different D2R data rate scenarios, where different transmission bandwidths and code rates can be used.
Observation 20: The performance of PDRCH with 32-bit preamble is close to that with 64-bit preamble, while the 16-bit preamble has a ~4dB performance loss.
Observation 21: The performance of PDRCH with a 4-bit preamble results in an error floor.
Observation 22: For PDRCH message sizes of 20 bits, inserting even a single midamble would cause an overhead of 22-29%, which is high when compared to the impact of overhead for other message sizes.
Observation 23: The overhead caused by the midambles for a 96-bit packet is less than 9%, but increases to 16% if two midambles are used.
Observation 24: For 8-bit midambles,
For a 400-bit packet, the addition of only 3 midambles result in an error floor at 1% BLER, but no error floor is seen when 4 midambles are used with an interval of ~28ms.
For a 96-bit packet, only 1 midamble is enough to avoid the error floor with an interval of ~30ms.
Observation 25: For 32-bit midambles,
For a 400-bit packet, the addition of only 2 midambles results in an error floor at 1% BLER, but no error floor is seen when 3 midambles are used with an interval of ~80ms. 
For a 96-bit packet, only 1 midamble is enough to avoid the error floor with an interval of ~90ms.
Observation 26: For 8-bit midambles and a 400-bit packet, the insertion of only 2 midambles result in an error floor at 1% BLER, but no error floor is seen when 3 midambles are used with an interval of ~28ms.
Observation27: For 32-bit midamble and a 400-bit packet, 7 midambles with an interval of ~42ms is required to avoid an error floor.

Proposal 1: For the R2D timing acquisition signal, the SIP is a dual ON-OFF transmission, which includes a detection threshold determination part (part #1), a high-voltage part (part #2) and a low-voltage part (part #3), which corresponds to Alt 2-4.
Proposal 2: The R2D SIP occupies a fixed duration of half of an OFDM symbol.
Proposal 3: The M value of the R2D SIP is M = 24 regardless of the M value of R2D CAP/PRDCH.
Proposal 4: The threshold determination part (part #1) of the R2D SIP has a duration that is 1/6th of an OFDM symbol duration, with each of the high- and low-voltage transmissions occupying 2 chips with M=24.
Proposal 5: The high-voltage part (part #2), or the ON part of the second ON-OFF transmission has a duration that is 1/12th of an OFDM symbol duration, occupying 2 chips with M=24.
Proposal 6: When determining the duration of the low-voltage part (part #3), or the OFF part of the second ON-OFF transmission, the following conditions have to be met:
Constant for different M values of M-chip OOK in order to achieve low complexity and reduce power consumption.
Required to satisfy the UHF RFID requirement, i.e., ≥ 12.5 μs.
Proposal 7: The low-voltage part (part #3), or the OFF part of the second ON-OFF transmission, has a duration that is 1/4th of an OFDM symbol duration, occupying 6 chips with M=24.
Proposal 8: For R2D SIP, our design can meet the MDR and FAR performance metric. 
Proposal 9: For the R2D SIP of the R2D timing acquisition signal, a dual ON-OFF transmission (Alt 2-4) is supported, with the following design aspects:
It contains 3 parts - a detection threshold determination part (part #1), a high-voltage part (part #2) and a low-voltage part (part #3).
It occupies a fixed duration of half of an OFDM symbol and the M value of the R2D SIP is M = 24 regardless of the M value of R2D CAP/PRDCH.
The detection threshold determination part (part #1) has a duration that is 1/6th of an OFDM symbol duration, with each of the high- and low-voltage transmissions occupying 2 chips with M=24.
The high-voltage part (part #2) has a duration that is 1/12th of an OFDM symbol duration, occupying 2 chips with M=24.
The low-voltage part (part #3) has a duration that is 1/4th of an OFDM symbol duration, occupying 6 chips with M=24.
No specific optimization is required to handle CP.
Proposal 10: The M value in CAP should be same as the M value in the subsequent PRDCH transmission.
Proposal 11: For the CAP of the R2D timing acquisition signal, its duration is variable (Option 1) for different M values.
Proposal 12: For the CAP of the R2D timing acquisition signal, it contains 4 chips and 2 transition edges in the same direction, has a fixed pattern, and begins with a high voltage.
Proposal 13: The maximum duration of the CAP of the R2D timing acquisition signal is 2 OFDM symbol, which is reached when M =2.
Proposal 14: For R2D time acquisition part, the required detection metric is the False Detection Ratio (FDR).
Proposal 15: For CAP, our design which keeps constant with the pattern of {high-, low-, high-, low-} voltages can meet the FDR performance metric.
Proposal 16: The M value set {2, 6, 24} is supported, and M = 4,8 is not be added due to an increased FDR.
Proposal 17: For the CAP of the R2D timing acquisition signal, a variable duration for different M values (Option 1) is supported, with the following design aspects:
Constant pattern of {high-, low-, high-, low-} is used for different M values.
Contains 4 chips and 2 transition edges in the same direction.
Begins with a high voltage.
Proposal 18: When designing the SIP, no specific optimization is required to handle CP.
Proposal 19: No specific optimization is required to handle CP for the CAP and the same CP handling method for PRDCH can be used.
Proposal 20: Device determines the end of R2D transmission based on violation of Manchester coding rule corresponding to the M value of the received PRDCH.
Proposal 21: D2R preamble is a single Golay or an m sequence with good correlation properties.
Proposal 22: For the D2R preamble, define two lengths of sequence:
An 8-bit short sequence for D2R transmissions when high data rate is expected to be achievable.
A 32-bit long sequence for D2R transmissions when lower data rate is expected to be achievable.
Note that there is no feedback regarding SNR. Similar as with disabling FEC, it can be determined by knowledge of the deployment.
Proposal 23: D2R midamble reuses the sequence in D2R preamble.
Proposal 24: D2R midamble uses two lengths of sequences, which is the same as that of the D2R preamble.
Proposal 25: D2R midambles are not used for shorter PDRCH transmissions with a packet size of 20 bits or less.
Proposal 26: For the interval between ambles in a D2R transmission, the maximum interval is 
~25ms for an 8-bit midamble.
~80ms for a 32-bit midamble.
Proposal 27: For the interval between the last midambles and the end of the D2R transmission, the maximum interval is 
~25ms for an 8-bit midamble.
~40ms for a 32-bit midamble.

3GPP TSG RAN WG1 #120bis		R1-2502275
Wuhan, China, April 7th – 11th, 2025

Source:	OPPO 
Title:                     Discussion on timing acquisition and synchronization for A-IoT
Agenda Item:       9.4.3
Document for:	Discussion and Decision

Conclusion
In this paper, we discussed the necessity and design of preamble/mid-amble/post-amble for R2D and D2R transmission in R19 A-IoT, we have following observations and proposals:

Observation 1: If the entire SIP duration is 1/2 or 1 OFDM symbol, Alt 1-1, Alt 1-2 with ON:OFF=1:2, and Alt 1-2 with ON:OFF=2:1 may not be differentiated from PRDCH transmission.
Observation 2: There is no need to always provide one part in every SIP for threshold detection training as this can be done by the device using the R2D transmission before the SIP.
Observation 3: Alt 1-2 with ON:OFF=1:3 can reach 1% MDR at 20dB SNR and also resistant to the impact of CP insertion. 

Observation 4: In contrast to post-amble, it is more reliable and efficient to indicate the end of PRDCH transmission with R2D control information. 
Observation 5: For the payload size of 20bits, preamble with 31bits provides no obvious performance gain but increase the overhead, compared to preamble with 15bit.
Observation 6: For the payload size of 96bits, preamble with 31bits achieve better performance at 1% BLER, compared to preamble with 64bits. 
Observation 7: Using preamble + midamble can achieve sufficient link performance for the payload size up to 400bits.
Proposal 1: SIP pattern should be different from other R2D transmissions.
Proposal 2: Multi-ON-OFF SIP pattern is not considered for the purpose of threshold detection training.
Proposal 3: Support Alt 1-2 with ON:OFF=1:3 and entire duration equal to 1/2 OFDM symbol as SIP for A-IoT.
Proposal 4: CAP design should ensure that CP insertion does not impact any chip length of CAP neither introduce false rising/falling edge.
Proposal 5: For clock-acquisition part, support Option 1 (i.e., variable duration for different M values).
Proposal 6: The end of PRDCH is indicated by R2D control information.
Proposal 7: Up to 3 relatively orthogonal sequences should be supported for preamble.
Proposal 8: M sequence is considered as the base sequence for preamble.
Proposal 9 Manchester coding should be applied to preamble, which is same as that for PDRCH.
Proposal 10 M-sequence lengths of 15bits and 31bits are supported for D2R preamble.
Proposal 11: Presence and location of D2R midamble are explicitly indicated by reader.
Proposal 12: No need to guarantee a midamble at the end of PDRCH transmission.

3GPP TSG RAN WG1 #120-bis								R1-2502319
Wuhan, China.  07-11 April 2025
Agenda Item 	:	9.4.3
Source 	:	Sony 
Title 	:	Timing acquisition and synchronisation for Ambient IoT 
Document for 	:	Discussion 

Conclusion 
This document has considered D-TAS and interference cancellation. The following proposal and observation are made:
Observation 1: Including an OFF period in the PDRCH preamble sequence allows for a shorter preamble. 

Proposal 1: For the purpose of interference mitigation of D2R, consider the option to include an OFF period known at the reader during the preamble transmission. 
 

3GPP TSG RAN WG1 #120bis		R1-2502370 
Wuhan, China, April 7th – 11th, 2025 
Agenda item:	9.4.3
Source:	Samsung
Title:	Views on timing acquisition and synchronization
Document for:	Discussion and decision
1	

Conclusion



In this contribution, we shared our views on timing acquisition and synchronization signals. More specifically, the discussion involves R2D/D2R preamble design, functionalities of midamble for D2R and the corresponding design, and functionalities of postamble for R2D/D2R and the corresponding design. To this end, the following observations and proposals were made:
In the previous sections we made the following observation:
Observation 1	In order for the SIP to be distinguishable from other part of R2D signal, the ratio of the ON duration to the OFF duration, denoted by X:1, can be chosen such that it cannot occur during the Manchester line coding. With Manchester encoding, X can be 1 or 2 at most. Therefore, X can be chosen to be greater than or equal to 3.
Observation 2	There is no need to specify a reference M value for the purpose of defining SIP. However, the fixed ON and OFF durations can be used by a device to obtain a certain level of chip synchronization by counting a number of samples from the known ON and OFF durations.
Observation 3	Postpone the determination of the maximum duration of CAP until the set of M values to be supported are decided. The down-selection shall be between 2 and 3 OFDM symbols, i.e., exclude 1.5 OFDM symbol to avoid complication in the CP handling.
Observation 4	Keeping constant duration of CAP for small M, i.e., Option 2, can be wasteful for small M, as the CAP will include an excessive number of chips.
Observation 5	For M ≥ 16, a chip duration becomes shorter than the CP, and the CP insertion can introduce a chip having an irregular length, which can affect the performance of a chip synchronization.
Observation 6	Since A-IoT D2R transmission is always triggered by a reader, the reader has a knowledge on when it is expected to be received. Also, PDRCH transmission is based on a simple OOK/BPSK modulation in time domain. Therefore, a sufficient level of synchronization performance can be achieved with a single part preamble signal.
Observation 7	In order to utilize postamble for synchronization and time tracking, and channel estimation for coherent detection, the receiver has to buffer the received signal and can start decoding only after reception of the postamble.
Observation 8	For R2D postamble, the functionality of channel estimation for coherent detection is not needed.  The functionality of synchronization and time tracking is also not needed as it can be achieved from PRDCH encoded with Manchester line code.
Observation 9	For D2R postamble, the functionalities of channel estimation for coherent detection, and synchronization and time tracking are not needed, since D2R midamble provides such functionalities.
Based on the discussion in the previous sections we propose the following:
Proposal 1	The SIP of R2D preamble is based on a single ON-OFF transmission having ON duration 3 times longer than OFF duration, i.e., Alt 1-3.
Proposal 2	Clarify whether inserted CP exists between the end of the SIP of R-TAS and the start of the CAP of R-TAS, and if yes, whether the inserted CP has a different voltage level compared to the end of the SIP of R-TAS and/or the start of the CAP of R-TAS.
Proposal 3	For CAP, adopt Option 1 (Duration of the clock-acquisition part is variable for different M values, i.e. the duration becomes shorter with increasing value of M).
Proposal 4	The CAP is based on a number of repetitions of ON-OFF patterns.
Proposal 5	The CAP duration becomes shorter with larger M, while the number of chips in the CAP increases with M.
Proposal 6	For M ≤ 12, the CAP is designed such that the last chip in each symbol duration has OFF state such that it does not incur any invalid chip in the CAP.
Proposal 7	For M ≥ 16, the CAP is designed such that the last chips copied as CP in each symbol duration have OFF state such that it does not incur any invalid chip in the CAP.
Proposal 8	D2R preamble is based on Golay sequence.
Proposal 9	For D2R preamble, support multiple sequences of same length or different length, where the sequence type is indicated in the preceding PRDCH or determined based on a predefined rule.
Proposal 10	D2R midamble reuses D2R preamble signal.
Proposal 11	The insertion of midamble can be based on 1) a predefined rule such as depending on the message type or size, or 2) based on explicit indication by a reader.
Proposal 12	The insertion of midamble can be every N number of chips of PDRCH, where N can be predefined or indicated to the device in the preceding PRDCH.
Proposal 13	The functionality of R2D postamble is only for determining the end of PRDCH.
Note: This does not preclude the use of TBS indication via L1 R2D control information, i.e., both R2D postamble and R2D control information can be used.
Proposal 14	The R2D postamble is based on single or multiple ON/OFF transmissions and distinguishable from other parts of the R2D transmissions.
Proposal 15	The functionality of D2R postamble is only for determining the end of PDRCH.

3GPP TSG RAN WG1 #120bis			R1-2502443
Wuhan, China, April 7th – 11th, 2025

Source:             Xiaomi
Title:                  Discussion on timing acquisition and synchronization for Ambient IoT
Agenda item:    9.4.3
Document for:  Decision

Conclusion
In this contribution, the following observations and proposals are provided:
Observations:
Observation 1: For the CP impact on SIP of R-TAS, it can be negligible.
Observation 2: For the CP impact on CAP of R-TAS, it is related to the M value.
When the M value is larger than or equal to 6, there is no impact on the timing precision of CAP.
When the M value is 2, the CP impact on the timing precision of CAP can be negligible due to the timing error is less than SFO.
Observation 3: When the packet size is 400bits, the midamble is essential to make an acceptable decoding performance, and 4 midambles is appropriate.
Observation 4: When the packet size is 96bits, only one midamble at the end of PDRCH is sufficient.
Observation 5:  Considering preamble and midamble of 32 bits, each PDRCH segment is not larger than ~100 bits when the total PDRCH is divided equally by midamble(s) and FEC is applied.

Proposals:
Proposal 1: For the SIP of R-TAS, Alt 1-2 (Single ON-OFF transmission) is preferred in which the OFF duration ratio is 1:3.
Proposal 2: For the CAP of R-TAS, Option 1 should be supported in which maximum duration is applicable to minimum value of M to be supported, and the CAP duration becomes shorter with increasing value of M.
The value of M should be consistent for the clock acquisition part and the subsequent PRDCH data part.
Proposal 3: For CAP of R-TAS, to fulfil with the agreed design principles, 
The value for maximum duration of CAP can be 1.5 OFDM symbol duration, and the CAP duration becomes shorter with increasing value of M.
The ON-OFF pattern can be fixed “ON-OFF-ON”.
Proposal 4: From the timing error perspective, the M value of 4 is not preferred for R2D transmission.
Proposal 5: To determine or derive the end of PRDCH transmission, Option 1 R2D postamble can be supported.
Option 1: R2D postamble immediately follows the PRDCH to indicate the end of the PRDCH.
Proposal 6: A recommended design of postamble will include contiguous high-level voltages longer than 2 R2D chips, e.g., 3 chips with contiguous high-level voltages.
Proposal 7: The D2R midamble can be omitted when the packet size of PDRCH is smaller than a threshold.
The threshold can be FFS.
Proposal 8: The sequence length and number of midamble should be configurable to achieve:
Flexible scheduling.
Relatively better decoding performance.
Relatively lower midamble overhead.
Proposal 9: For the D2R preamble which is immediately transmitted before any PDRCH transmission, the following two parts can be considered:
Part 1: pilot of all one or all zero with square waves or Manchester coding.
Part 2: binary sequence with good correlation performance. 
Proposal 10: For the binary sequence of D2R preamble with good correlation performance, Golay or m sequence can be supported.
Proposal 11: D2R midamble can be an opposite sequence as the part 2 of preamble per bit.
Proposal 12: For a specific sequence length of preamble or midamble, an Golay or m sequence can be supported.
Proposal 13: To achieve sufficient scheduling flexibility, multiple sequence lengths of preamble and midamble can also be supported.

3GPP TSG RAN WG1 #120bis			R1-2502468
Wuhan, China, 7th – 11th April 2025

Source:	Panasonic
Title: 	A-IoT Timing Acquisition and Synchronization
Agenda Item:		9.4.3
Document for:	Discussion, Decision

Conclusion 
In this contribution, the followings proposals are made: 
Observation 1: The threshold setting affects the evaluation accuracies. Since the threshold might be easily influenced by larger ON/OFF ratio, the ratio of = 1:3 or 3:1 should be deprioritized.
Observation 2: Equal length of ON/OFF ratio is insensitive to the thresholds setting.

Proposal 1: Multi-ON-OFF transmission (Alt 2) can be used. The first instance of ON and OFF is used for AGC setting and threshold setting.
Proposal 2: Consider Alt 2-1 as the repetition of Alt 1-1 as the starting point. To avoid the similarity with Manchester coding and CAP should be further discussed.
Proposal 3: For CAP of R-TAS from TR 38.769, the minimum requirement is Option 1. If the complexity is not increase too much, it’s also ok to support Option 2.
Proposal 4: Different device type does not share a transaction start from start indicator, clock acquisition part, Msg 0, Msg 1 and so on. The mechanism of "the transaction for device 1 is not triggered by the device 2a/2b" and "the transaction for device 2a/2b is not triggered by the device 1" should be supported. Whether the different device type can share the same frequency with transaction level should be concluded later.
Proposal 5: The cost, power consumption and coverage trade-off should be achieved for each of device 1 and device 2a/2b instead of too much focus on the common design among device 1/2a/2b as far as the transaction level compatibility is achieved.

3GPP TSG RAN WG1 #120bis                                                         R1-2502471		                                                   
Wuhan, China, April 7th – 11th, 2025

Agenda item:   9.4.3
Source: 	    Lekha Wireless Solutions
Title:	               Timing acquisition and synchronization for Ambient IoT

Document for:    Decision



1. 

Conclusion
In this document, different options for the sequences for preamble design for the Ambient IoT devices are discussed and the following observations and proposals are made:

Observation 1: m-sequences (maximum-length sequences) or PN sequences are binary sequences generated by LFSRs that appear random but are deterministically produced. They have a maximum length of 2^n-1 for an n-stage LFSR and are balanced with equal 1s and 0s. Autocorrelation of m-sequences results in a peak value at zero shift and near-zero values (-1/N) at other shifts. The cross-correlation plot generated by different polynomials shows high peaks at certain lags, which results in m-sequences causing multiple access interference and leading to performance degradation.

Observation 2: Golay sequences are a pair of complementary binary sequences with ideal autocorrelation properties. When combined, their autocorrelation sum is zero, making them highly effective in reducing interference.
Observation 3: Gold sequences are binary sequences generated by combining two preferred m-sequences using modulo-2 addition.  The length of a Gold sequence is typically N=2^n-1, where n is the LFSR stage count. A total of N+2 sequences can be generated by combining two m sequences. They have low cross-correlation between different sequences, reducing interference among multiple devices. Their autocorrelation has a peak at zero shift and low values at other shifts.
Observation 4: Zadoff-Chu sequences are complex-valued sequences with constant amplitude, making them power-efficient. They have ideal autocorrelation properties, with a peak at zero shift and 0 value at other shifts. These sequences have good cross correlation properties, minimizing interference. Their length is typically a prime number, and they are generated based on a root. They are widely used in modern systems like LTE and 5G for synchronization and reference signals.

Proposal 1: Gold sequences are incredibly effective in minimizing interference and ensuring reliable communication, making them a good choice for multiple access systems. They have excellent autocorrelation and cross correlation properties. Gold sequences offer flexibility and scalability, as they can be easily generated by XORing preferred m sequences. Hence, Gold sequences are supported for D2R preamble as they provide a robust solution for high-performance, interference-resistant communication.

4. 

3GPP TSG RAN WG1 #120bis			                 R1-2502476
Wuhan, China, April 7th – 11th, 2025

Agenda Item:	9.4.3
Source: 	LG Electronics
Title: 	Timing acquisition and synchronization for A-IoT
Document for:	Discussion and decision

Conclusions
In this contribution, we discussed the timing acquisition and synchronization for Rel-19 Ambient IoT, we have the following proposals and observation.

Proposal #1: Support the followings for start-indicator part (SIP) of R2D timing acquisition signal (R-TAS). 
Entire duration: 1 OFDM symbol
Pattern: {OFF, ON, OFF}
The last OFF duration are defined as more than twice the ON duration. For example,
If chip duration of SIP is M=8, {OFF, ON, OFF} = {1, 2, 5} chips.
If chip duration of SIP is M=12, {OFF, ON, OFF} = {1, 3, 8} chips.
If chip duration of SIP is M=16, {OFF, ON, OFF} = {3, 4, 9} chips.
FFS whether to include CP length when defining the chip duration of SIP.

Observation #1: If maximum duration of clock-acquisition part (CAP) is 1.5 OFDM symbol, the starting point of PRDCH cannot be the OFDM symbol boundary. Also, if maximum duration of clock-acquisition part (CAP) is 3 OFDM symbols, it can increase the signaling overhead.

Proposal #2: Support the followings for clock-acquisition part (CAP) of R2D timing acquisition signal (R-TAS). 
Maximum duration: 2 OFDM symbols
Pattern: {ON, OFF, ON, OFF, ON}
For example, assuming the M value for PRDCH is 1 and the CAP design is based on M = 8, the chip duration of CAP can be defined as {ON, OFF, ON, OFF, ON} = {1, 5, 3, 5, 2} chips.
FFS whether to include CP length when defining the chip duration of CAP.

Proposal #3: Support hybrid option (Option 1 + Option 2) for Clock-acquisition part (CAP) for R2D timing acquisition signal (R-TAS).
CAP length would be reduced as M increases when M is less than MTH, and CAP length would be maintained using repetition when M is greater than or equal to MTH.
Define a chip duration of PRDCH as the time interval of two adjacent rising (or falling) edges of the CAP.
Align an ending time of CAP with the OFDM symbol boundary.
Further discuss whether to support providing additional information (e.g., M value indication, release number, device type, etc.) by changing the ON/OFF pattern of the repeated CAP.

Proposal #4: Discuss whether/how to apply padding between the end of CAP and the start of PRDCH pending the decision on the CP handling method.

Proposal #5: Support R2D TBS information indication via explicit L1 R2D control information as baseline operation (Option 1)
Further discuss whether to additionally support R2D postamble

Proposal #6: M-sequence can be considered for D2R timing acquisition signal (D-TAS)

Proposal #7: At least two formats (e.g., normal format and extended format) can be considered for D2R timing acquisition signal (D-TAS)

Proposal #8: Support D2R TBS information indication via explicit L1 R2D control information

Proposal #9: If D2R midamble is supported for D2R transmission, introduce a time duration and/or a payload size based on which device determines whether and how often D2R midamble should be transmitted.
Further discuss whether to transmit D2R midamble depending on the modulation of D2R transmission.

Proposal #10: Support the reader can indicate whether to include D2R midamble in the D2R transmission via R2D control information.

Proposal #11: If D2R midamble is supported for D2R transmission, it should be distinguished from the pattern(s) in D-TAS.

3GPP TSG RAN WG1 #120-bis			R1-2502511
Wuhan, China, April 7th – 11st, 2025

Source:	ETRI
Title:	Discussion on timing acquisition and synchronization
Agenda item:	9.4.3
Document for:	Discussion

Conclusion
In this contribution, we have discussed on the aspects of timing acquisition and synchronization signals and provide following observations and proposals.

Observation 1: The SIP pattern should be clearly distinguishable from all possible OOK-M chip patterns when Manchester encoding is applied.

Observation 2: It is recommended that the SIP pattern includes a continuous ON or OFF period that is at least three times the chip duration for M = 2. 

Observation 3: If M = 1 would be included among the supported M values for the OOK of R2D PRDCH, it is required to set the CAP duration to two OFDM symbols in R-TAS. 

Observation 4: Use of R2D control information may cause unnecessary control overhead. 

Observation 5: If applications such as sensors are considered in a future release, the length of R2D transmission may exceed 1000 bits which may cause backward-compatibility issue for R2D control information.

Observation 6: R2D control information does not provide any additional robustness gain compared to the postamble due to the absence of FEC for R2D. 

Proposal 1: For entire duration of SIP, one OFDM symbol duration is supported. 

Proposal 2: For a pattern of SIP, Alt 1-3: ON followed by OFF with a duration ratio of 3:1 is supported.

Proposal 3: For the duration of CAP, the only (constant) duration is supported for all the M values (Option 2). 

Proposal 4: When the minimum M value is 1, a CAP duration of two OFDM symbols is proposed as the only (constant) duration. 

Proposal 5: To address the CP issue, it is proposed that the remaining portion—after configuring the basic(default) patterns near the OFDM symbol boundary (for M ≥ 2) or CAP boundary (for M = 1)—be set to OFF (low voltage). 

Proposal 6: It is proposed to first determine the minimum M value to be supported. 

Proposal 7: If M = 2 is supported as the minimum supported M value, a 1-OFDM-symbol duration may be also considered as the only (constant) CAP duration for the down-selection. 

Proposal 8: To determine or derive the end of PRDCH transmission, R2D postamble immediately follows the PRDCH to indicate the end of the PRDCH is supported. 

Proposal 9: If the end of R2D transmission based on violation of Manchester coding rule would be supported, it should ensure that reader does not utilize SIP pattern as a violation part. 

Proposal 10: It is proposed that, for Rel-19, the reader explicitly indicates the midamble(s) through R2D control information.

3GPP TSG RAN WG1 #120bis		        R1-2502607
Wuhan, China, April 7th – 11th, 2025

Agenda Item:	9.4.3
Source:	Apple
Title:	On timing acquisition & synchronization aspects for Ambient IoT
Document for:	Discussion/Decision

Conclusion
In this contribution, following observations/proposals have been made on R2D and D2R signals:

R2D Signals

Observation 1: With the assumption that there is no RF transmission available before the SIP, therefore for reliable detection of SIP, some threshold detection training is needed.



Observation 2: If the end of SIP, based on Alt 2-4 is OFF, then the start of CAP has to be ON and to include then two rising edges, ON-OFF-ON pattern can be sufficient.
1st rising edge is transition from the end of SIP to start of CAP, i.e. from OFF to ON
2nd rising edge is transition from second to third chip of CAP, i.e. from OFF to ON

Observation 3: If CAP spans more than 1 OFDM symbol duration, then M > 12 cannot be reliability indicated as the duration of CP becomes longer than the CAP chip duration

Observation 4: Following observations can be made regarding the alternatives for indicating the end of PRDCH transmission:
Manchester-coding rule violation-based alternatives
Alt 1: Postamble pattern is specified to be attached at the end of PRDCH that violates the rule
This would require further discussion on exact design based on different options from companies
Alt 2: Postamble pattern is not specified, and reader ensure violation
This wouldn’t require any further discussion on postamble design
R2D control information-based alternatives
Alt 3: TBS via R2D control information with separate CRC attachment 
This may require further discussion on potentially different PRDCH formats/designs
Alt 4: TBS via R2D control information with joint CRC attachment
This may require early detection of R2D control information without CRC check, but may impact the reliability of detection




Proposal 1: For SIP of R-TAS, adopt Alt 2-4 with a SIP patter including ON-OFF-ON-OFF-OFF-OFF

Proposal 2: Overall duration of R-TAS occupies entire duration of one or more OFDM symbols rather than partial duration.

Proposal 3: For the CAP of R-TAS, ON-OFF-ON pattern can be adopted consider Alt 2-4 is adopted for SIP

Proposal 4: For the CAP of R-TAS, adopt option 2, i.e. duration of the clock-acquisition part is constant for different M values based on repetition, i.e. repetition factor is increased with increasing value of M to keep the duration constant.

Proposal 5: CAP pattern occupies a fixed duration of 1.5 OFDM considering three chips and supporting minimum value of M = 2

Proposal 6: To indicate M > 12 for PRDCH, adopt one of the following mapping options between M value used for CAP and corresponding M value determined for CAP
Option 2-1: If M = 2, 6, 12 and 24 are adopted for PRDCH, corresponding M values of CAP include M = 2, 6, 8, 12, respectively
Option 2-2: If M = 2, 8, 16 and 24 are adopted for PRDCH, corresponding M values of CAP include M = 2, 6, 8, 12, respectively

Proposal 7: Adopt Alt 2, i.e. postamble pattern is not specified and reader ensures Manchester coding rule violation for indicating the end of PRDCH transmission
No further discussion needed on details of any postamble pattern/design


D2R Signals

Observation 5: If two preamble types including long preamble format and short preamble format are adopted for D2R, then during the random access procedure as the reader may send paging message to multiple devices and those devices may have different channel conditions, therefore, it may be difficult for reader to determine which preamble type is suitable



Proposal 8: For D2R x-ambles, base Sequence type, as there is no significant performance different between M-sequence and Golay-sequence and therefore one of the sequences can be adopted based on majority view

Proposal 9: For D2R preamble, a long and a short preamble format can be supported

Proposal 10: If two preamble types including long preamble format and short preamble format are adopted for D2R, then default format is long preamble format for the device to apply for Msg1

Proposal 11: If two preamble types including long preamble format and short preamble format are adopted for D2R, then after the random access procedure is concluded, reader should be able to indicate which format type is used by the device.

Proposal 12: D2R preamble is always present and immediately precedes PDRCH

Proposal 13: D2R midamble uses the same sequence as D2R preamble as same functionality is served by both 

Proposal 14: Combination of pre-defined rule and reader indication can be used to indicate the presence and number of midambles and their corresponding time-domain resources 
In the absence of reader indication, pre-defined rule is applied

Proposal 15: Joint signaling of preamble and midamble configuration is supported indicating preamble type (if multiple types supported), presence of midamble, number of midambles and time-domain resources/location of midamble

3GPP TSG RAN WG1 #120bis		        					     R1-2502609
Wuhan, China, April 7th – 11th, 2025

Agenda Item:	9.4.3
Source: 	Moderator (Apple)
Title:	FL Summary#1 on timing acquisition & synchronization for Ambient IoT
Document for:	Discussion & Decision

Introduction & Work-Plan for RAN1#120bis
This document provides the feature lead summary on the offline discussions/inputs/proposals for AI 9.4.3 timing acquisition and synchronization signals for R2D and D2R for ambient IoT WI during RAN1#120bis. 


Contact Information
Please consider providing your company name, your name and email address to be able to reach for any potential offline discussions/contact regarding AI 9.4.3 on timing acquisition and synchronization for ambient IoT.





Work Plan
Following is my plan for this meeting for this agenda for RAN1#120bis:

SIP of R-TAS
For SIP, at least down-select to not more than 2 alternatives from the agreed alternatives in RAN1#120 considering technical justification and evaluation analysis. Also agree on fixed single duration for the down-selected alternatives. 
CAP of R-TAS
For CAP, plan is to down-select between option 1 and option 2 (as agreed in last meeting) and then also limit the number of specific CAP pattern design and hopefully have not more than 2 patterns for further down-selection by end of this meeting
R2D Postamble aspects
For indicating the end of PRDCH transmission, at least down-select to whether R2D control information is used, or Manchester coding rule violation is used. 
D2R X-amble
For D2R x-amble, first priority is to agree on the sequence type and length(s). In addition, for midambles, agree on the method/signaling that allows the device to determine the number of midamble(s) and their location within the PDRCH.

SIP of R-TAS

[Closed] 1st Discussion Round 
For the SIP of R-TAS, following aspects have been discussed in the contributions following the agreements and discussion from RAN1#120:
Preferred alternative(s) for SIP design
Duration of SIP
For comparison among different alternatives, companies also provided simulations results/analysis to justify their preferred alternative(s). In addition, a few companies also discussed other aspects including M value/chip duration for the SIP, whether/what impact of CP insertion on SIP. In the following table, companies’ views on their preferred alternatives along with justification and performance analysis, if provided are summarized. 

Table 2-1: Summary of views on alternatives for SIP



FL Observations:
28 companies provided their views on SIP of R-TAS
11 companies provided evaluation in terms of target MDR and FAR with most of the companies considering MDR <= 1% for a target FAR <= 1%
For Alt 1-1, 3 companies provided their preference with ON-OFF ratio of 1:1 & total duration spanning 1 OFDM symbol and only 1 company [China Telecom] showed that it performs better than other alternatives while it satisfies the target MDR and FAR. On the other hand, 4 companies [Futurewei, Xiaomi, Samsung, Huawei] showed other alternatives that perform better than Alt 1-1.
For Alt 1-2, overall, 8 companies provided their preference for at least some variant of Alt 1-2 and 6 companies [Futurewei, ZTE, Spreadtrum, Oppo, Xiaomi, Qualcomm] showed that at least some variant of Alt 1-2 performs better than other alternatives including Alt 1-1, Alt 1-3, Alt 2-2, Alt 2-4. On the other hand, 4 companies [China Telecom, Samsung, CMCC, Huawei] showed other alternatives that performs better than Alt 1-2. Within Alt 1-2:
2 companies prefer ON-OFF with ratio of 1:3 & total duration spanning 0.5 OFDM symbol
5 companies prefer ON-OFF with ratio of 1:3 & total duration spanning 1 OFDM symbol
3 companies prefer ON-OFF with ratio of 1:2 & total duration spanning 1 OFDM symbol
For Alt 1-3, overall, 6 companies provided their preference for at least some variant of Alt 1-3 and only 1 company [Samsung] showed that at least some variant of Alt 1-3 performs better than other alternatives including Alt 1-1, Alt 1-2. On the other hand, 6 companies [China Telecom, ZTE, Futurewei, Oppo, Qualcomm, Huawei] showed other alternatives that performs better than Alt 1-3. Within Alt 1-3:
4 companies prefer ON-OFF with ratio of 2:1 & total duration spanning 1 OFDM symbol
5 companies prefer ON-OFF with ratio of 3:1 & total duration spanning 1 OFDM symbol
For Alt 2-1, 7 companies provided their preference with ON-OFF ratio of 1:1 & total duration spanning 1 OFDM symbol and no company showed that it performs better than other alternatives. On the other hand, 2 companies [China Telecom, Vivo] showed other alternatives that perform better than Alt 2-1.
For Alt 2-2, overall, 6 companies provided their preference for at least some variant of Alt 2-2 and only 1 company [Vivo] showed that at least some variant of Alt 2-2 performs better than other alternatives including Alt 2-1, Alt 2-4. On the other hand, 2 companies [China Telecom, ZTE] showed other alternatives that performs better than Alt 2-2. Within Alt 2-2:
1 company prefer ON-OFF-ON with ratio of 1:1:4 & total duration spanning 1 OFDM symbol
2 companies prefer ON-OFF-ON with ratio of 1:4:1 & total duration spanning 1 OFDM symbol
3 companies prefer total duration spanning 1 OFDM symbol without any specific consideration on ration of ON/OFF
For Alt 2-3, overall, 3 companies provided their preference for at least some variant of Alt 2-3 and no company showed that it performs better than other alternatives including. On the other hand, 2 companies [China Telecom, Huawei] showed other alternatives that performs better than Alt 2-3. Within Alt 2-3:
2 companies prefer OFF-ON-OFF with ratio of 1:1:4 & total duration spanning 1 OFDM symbol
1 company prefer OFF-ON-OFF with ratio of 1:4:1 & total duration spanning 1 OFDM symbol
1 company prefer OFF-ON-OFF with total duration spanning 1 OFDM symbol
For Alt 2-4, 4 companies provided their preference with ON-OFF-ON-OFF ratio of 1:1:1:3 & total duration spanning 0.5 OFDM symbol and 2 companies [CMCC, Huawei] showed that it performs better than other alternatives. On the other hand, 2 companies [ZTE, Vivo] showed other alternatives that perform better than Alt 2-4.
A few companies discussed M value for SIP and whether/how the selection of M may impact the SIP detection considering CP insertion. However, based on discussions during RAN1#120, we considered discussing in terms of OFDM symbol duration and avoid reference M value. Therefore, with that understating, we should not try to reopen the discussion of reference M value but rather have discussion in terms of OFDM symbol duration. Eventually, it will be up to the specification editor on how to capture the SIP duration and corresponding ON(s) and OFF(s). Furthermore, at least 1 company raised the issue that for chip duration corresponding to M < 6, minimum PRB of 2 will be needed. However, another company clarified that we don’t need to associate the mapping of M to minimum number of PRBs for SIP and device should be able to adjust its LPF between SIP reception and follow-up R2D reception. 
Another aspect that a few discussed is whether/what impact CP might have on SIP detection. One company pointed out that chip duration corresponding to higher value of M map impact SIP detection, while 3 companies pointed out that CP impact is negligible on SIP and should not be considered for specific optimization, especially considering that all companies considered SIP duration of not more than 1 OFDM symbol

Based on above observations from companies’ contributions, as a starting point, FL proposal would be to down-select to Alt 1-2 and Alt 2-4. Main rationale behind these two alternatives is not just majority preference but evaluations showing that these two alternatives at least perform better within alternative set 1 and set 2, respectively. Further focused discussion on Alt 1-2 and Alt 2-4 can be further done, if we down-select these 2 options. Therefore, accordingly Proposal 2-1 is provided below:

Proposal 2-1 
For the pattern of SIP of R-TAS, only following 2 alternatives are considered for further down-selection to one alternative:
Alt 1-2: ON-OFF with a ratio of 1:3 and with following total SIP duration to be further down-selected during RAN1#120bis:
Option 1: 0.5 OFDM symbol duration
Option 2: 1 OFDM symbol duration
Alt 2-4: ON-OFF-ON-OFF with a ratio of 1:1:1:3








CAP of R-TAS


[Closed] 1st Discussion Round 
For the CAP of R-TAS, following aspects have been discussed in the contributions following the agreements and discussion from RAN1#120:
Preferred option in terms of variable or fixed duration for CAP design
Preferred exact CAP pattern including M values/duration and CP insertion related aspects

Table 3-1: Summary of views between option 1 and option 2 for CAP




Table 3-2: Summary of views on exact CAP pattern


FL observations
Overall, 30 companies provided their views on CAP of R-TAS
26 companies provided their preference relation to option 1 and option 2 from the TR and 17 companies indicated their preference for option 1, while 8 companies indicated their preference for option 2. 1 company indicated the preference to support and combine both options. Based on the agreement from RAN1#120bis, one option needs to be down-selected in this meeting. From the proponents of option 1, 3 companies provided evaluations and showed that option 1 is able to achieve target FDR, as long as there is sufficient gap between different M values. On the other hand, from the proponents of option 2, 2 companies provided evaluations and showed that at least for higher M values, option 2 is needed. 
Among specific CAP patterns, 5 alternatives are provided. Among the 5 alternatives, majority shows that either 3 chip with alternate ON/OFF or 4 chips with alternate ON/OFF can work well for the M values. Alt 3, 4, 5 with more chips have been proposed only by 1-2 companies each and don’t offer any significant gain compared to Alt 1 and Alt 2 and rather increase the overhead, especially for lower M values
Based on above, proposal 3-1 and proposal 3-2 are provided.



Proposal 3-1
For CAP of R-TAS, option 1 for CAP of R-TAS from TR 38.769 is adopted with following details:
CAP duration becomes shorter with increasing value of M




Proposal 3-2 
For the pattern of CAP of R-TAS, only following 2 alternatives are considered for further down-selection to one alternative:
Alt 1: 3 chips with ON-OFF-ON pattern 
Alt 2: 4 chips with ON-OFF-ON-OFF pattern
Based on the minimum value of M that will be agreed in agenda 9.4.2, corresponding maximum duration of CAP will be selected from the candidate set of {1.5 OFDM symbols, 2 OFDM symbols, 3 OFDM symbols}




R2D Postamble



[Closed] 1st Discussion Round 
FL observations
24 companies provided their views in terms of different options for indicating the end of PRDCH transmission
Overall, 11 companies prefer indicating the end of PRDCH transmission based on some unique pattern that violate Manchester coding rule:
7 companies prefer to explicitly specify postamble design
4 companies prefer to not specify postamble design and leave it up to the reader implementation to violate the Manchester coding rule
8 companies prefer indication the end of PRDCH transmission based on R2D control information. Among the 8 companies there are still different views on whether R2D control information is L1 or not
3 companies consider supporting combination of the above 2 options
From FL perspective, all the options work and doesn’t necessarily have any significant performance difference. However, for R2D control indication based method, it is still not clear whether L1 R2D control information is used or not. Based on the proponents, the views are still diverging. If L1 R2D control information is used, then it needs further discussion on separate or joint CRC attachment. On the other hand, for Manchester coding rule based option, in principle companies have common understanding that violation, for example could be based on transmission of three high or three low voltages, i.e. ON-ON-ON or OFF-OFF-OFF. But the main divergence is whether it needs to be specified or not.


Based on above, proposal 4-1 is provided to take the majority view of indicating end of PRDCH transmission based on violation of Manchester coding rule and also with the consideration that with this option, no further discussion is needed on details related to R2D control information and CRC. For method based on violation of Manchester coding rule, furthermore, it can be discussed whether/what postamble needs to be specified or not. If companies cannot converge whether/what postamble needs to be specified or not, then implementation based solution can be adopted.


Proposal 4-1 

For indicating the end of PRDCH transmission, Manchester coding rule violation corresponding to M value is adopted
FFS: whether the Manchester coding rule violation is based on specified postamble pattern or based on device detecting such violation based on reader’s implementation
Note: If companies cannot converge on whether/what postamble is specified or not, then device detecting such violation based on reader’s implementation will be adopted
Note: If no postamble is specified, then from timing perspective, end of last chip of PRDCH is the end of R2D transmission




D2R X-ambles (including preamble & midamble)


[Closed] 1st Discussion Round 
Base Sequence Types: First aspect discussed by companies is related to the preference between the two sequence types including the length that were agreed in RAN1#120bis to be further down-selected and corresponding details including sequence length and generation. In Table 5-1 below, the companies’ views are summarized and also including performance analysis, if provided

Table 5-1: Summary of views between M-sequence and Golay Sequence types for D2R X-ambles



Multiple Preamble lengths: Second aspect discussed by companies is related to the number of preamble types in terms of different lengths. In Table 5-2 below, the companies’ views are summarized and also including performance analysis, if provided

Table 5-2: Summary of views on D2R preamble lengths including long and short preamble



# of Preamble Sequences (with same base sequence): Third aspect discussed by companies is related to the number of sequences of same lengths for preamble/midamble. In Table 5-3 below, the companies’ views are summarized and also including performance analysis, if provided

Table 5-3: Summary of views on number of preamble sequences



1- or 2- part Preamble: Fourth aspect is related to 1-part or 2-part preamble design. In Table 5-4 below, the companies’ views are summarized and also including performance analysis, if provided

Table 5-4: Summary of views on number of parts of preamble



Midamble Configuration: Fifth aspect is related to presence/location of D2R x-ambles including whether pre-defined rule and/or explicit signaling by reader is supported. In Table 5-5 below, the companies’ views are summarized and also including performance analysis, if provided

Table 5-5: Summary of views on midamble configuration





Midamble Sequence: Fifth aspect is related to whether midamble sequence is same as preamble or not. In Table 5-6 below, the companies’ views are summarized and also including performance analysis, if provided

Table 5-6: Summary of views on midamble sequence



FL observations
In terms of the base sequence type, all the companies that provided simulation results showed that the performance is similar with both Golay sequence and M sequence. And in terms of preference, almost there is equal number of companies supporting both options. For M sequence, it was pointed out by few companies that the sequence generation is perhaps simpler compared to Golay sequence. Therefore, FL proposal is to adopt M sequence for D2R x-ambles. 

In terms of preamble types, at least 11 companies prefer to support two preamble formats including short preamble format and long preamble format. Multiple companies pointed out the benefit of supporting the two formats and also, one company demonstrated the benefit of using two lengths. In terms of sequence length, 32(-1) bits is proposed by majority of company. For the shorter format, it was shown that 8(-1) bits have reasonable performance difference, while 16-(1) bits have almost similar performance. Therefore, FL proposes to support both short and long preamble format, with length 8(-1) bits and 32(-1) bits, respectively. Also, all but 2 companies considered same sequence for both midamble and preamble as both serve similar functionality and this would require less specification effort. Therefore, same sequence for both is also proposed by FL. In terms of 1-part or 2-part format for preamble, 4 companies prefer to have 2-part for coarse and fine synchronization, respectively. On the other hand, 2 companies explicitly proposed not to support 2-part preamble. One company showed in evaluations that as long as preamble and midamble are designed optimally in terms of length and sequence, there is not benefit from 2-part preamble. In order to get views from more companies, FL will ask companies to provide inputs on their preference

Regarding the configuration of midambles, ~13 companies prefer explicit indication by the network, while 8 companies prefer pre-defined rule to determine number of midambles and location. In terms of pre-defined rule, majority consider supporting TBS based midamble determination. Furthermore, there are 7 companies prefer a combination of pre-defined rule and explicit indication by network. From FL perspective, all solutions work, but it seems there is a quite good number of companies that prefer additional control to reader in terms of at least the presence and number of midamble due to potentially dynamic change in channel conditions and/or reader-specific implementation. Therefore, FL proposal would be to consider the combined option where the number of midambles can be explicitly indicated by the reader to the device via R2D control information and the position of midambles can be at least pre-defined. 


Based on above summary of views from companies on D2R x-ambles, following proposals are provided




Proposal 5-1 
For D2R preamble and midamble, base sequence is generated from M-sequence, where the length of the sequence is 
Value(s) of n = 5, 3
Long preamble is generated based on n = 5
Short preamble is generated based on
 n = 3

FFS: signaling/application of the two preambles
Same base sequence is applied for both preamble and midamble
Preamble immediately precedes the PDRCH without any gap


Question 5-1 
Do you support 1-part or 2-part preamble, where for 2-part preamble, 1st part is based on 0s on 1s without Manchester coding and 2nd part is based on the base sequence (as proposed in proposal 5-1)?


Proposal 5-2
For D2R midamble, following for the determining the presence and location of midamble(s) at the device:
Reader explicitly indicates the number of midambles to be applied by the device for PDRCH transmission via R2D control information
FFS: details of signaling
Based on the indicated number of midambles, a pre-defined rule is applied at the device to determine the exact location of each of the midamble(s) 
FFS: details of pre-defined rule for determining the exact location of midamble(s) 
















Proposals for offline sessions
1st offline session (Monday, April 7, 2025)
(Updated) Proposal 5-2
For D2R midamble, following two alternatives are considered for further down-selection in RAN1#120bis for the determining the presence and location of midamble(s) at the device:
Alt 1: Reader explicitly indicates the number of midambles to be applied by the device for PDRCH transmission via R2D control information
FFS: details of exact signaling to be discussed under agenda 9.4.4
Based on the indicated number of midambles, pre-defined rule(s) are applied at the device to determine the exact location of each of the midamble(s) 
FFS: details of pre-defined rule for determining the exact location of midamble(s)
Alt 2: Reader explicitly indicates the interval between consecutive x-ambles
FFS: details of exact signaling
Note: These alternatives don’t preclude support of no midamble

(Updated) Proposal 3-1
For CAP of R-TAS, option 1 for CAP of R-TAS from TR 38.769 is adopted with following details:
CAP duration becomes shorter with increasing value of M
FFS: Exact relation between duration of CAP and M values

(Updated) Proposal 2-1 
For the pattern of SIP of R-TAS, only following 2 alternatives are considered for further down-selection to one alternative:
Alt 1-2: ON-OFF with a ratio of 1:3 and with following total SIP duration to be further down-selected during RAN1#120bis:
Option 1: 0.5 OFDM symbol duration
Option 2: 1 OFDM symbol duration
Alt 2-4: ON-OFF-ON-OFF with a ratio of 1:1:1:3 and with following total SIP duration to be further down-selected during RAN1#120bis:
Option 1: 0.5 OFDM symbol duration
Option 2: 1 OFDM symbol duration

Proposal 4-1 
For indicating the end of PRDCH transmission, Manchester coding rule violation corresponding to M value is adopted
FFS: whether the Manchester coding rule violation is based on specified postamble pattern or based on device detecting such violation based on reader’s implementation
Note: If companies cannot converge on whether/what postamble is specified or not, then device detecting such violation based on reader’s implementation will be adopted
Note: If no postamble is specified, then from timing perspective, end of last chip of PRDCH is the end of R2D transmission


(Updated) Proposal 5-1 
For D2R, 1-part preamble and 1-part midamble, base sequence is generated from M-sequence, where the length of the sequence is 
Value(s) of n = 5, 3
Long preamble is generated based on n = 5
Short preamble is generated based on following n values to be down-selected to single value
Alt 1: n = 3
Alt 2: n = 4
FFS: signaling/application of the two preambles
Same base sequence is applied for both preamble and midamble
Preamble immediately precedes the PDRCH without any gap

Proposal 3-2 
For the pattern of CAP of R-TAS, only following 2 alternatives are considered for further down-selection to one alternative:
Alt 1: 3 chips with ON-OFF-ON pattern 
Alt 2: 4 chips with ON-OFF-ON-OFF pattern
Based on the minimum value of M that will be agreed in agenda 9.4.1, corresponding maximum duration of CAP will be selected from the candidate set of {1.5 OFDM symbols, 2 OFDM symbols, 3 OFDM symbols}






























Proposals for online session
1st  online session (Tuesday, April 8, 2025)
Proposal 5-2a
For D2R midamble, following two alternatives are considered for further down-selection in RAN1#120bis for determining the presence and location of midamble(s) at the device:
Alt 1: Reader explicitly indicates the number of midambles to be applied by the device for PDRCH transmission via R2D control information
FFS: details of signaling
Based on the indicated number of midambles, pre-defined rule(s) are applied at the device to determine the exact location of each of the midamble(s) 
FFS: details of pre-defined rule for determining the exact location of midamble(s) 
Alt 2: Reader explicitly indicates the interval between consecutive x-ambles
FFS: details of signaling and any pre-define rule(s), if needed
Note: These alternatives don’t preclude the support of no midamble


Proposal 2-1a
For the pattern of SIP of R-TAS, only following 2 alternatives are considered for further down-selection to one alternative:
Alt 1-2: ON-OFF with a ratio of 1:3 and with following total SIP duration to be further down-selected [during RAN1#120bis]:
Option 1: 0.5 OFDM symbol duration
Note: In this option, the chip duration corresponds to M = 8, but this doesn’t imply support of such reference M value for SIP
Option 2: 1 OFDM symbol duration
Note: In this option, the chip duration corresponds to M = 4, but this doesn’t imply support of such reference M value for SIP
Alt 2-4: ON-OFF-ON-OFF with a ratio of 1:1:1:3 and with following total SIP duration to be further down-selected [during RAN1#120bis]:
Option 1: 0.5 OFDM symbol duration
Note: In this option, the chip duration corresponds to M = 12, but this doesn’t imply support of such reference M value for SIP
Option 2: 1 OFDM symbol duration
Note: In this option, the chip duration corresponds to M = 6, but this doesn’t imply support of such reference M value for SIP

Proposal 3-1a
For CAP of R-TAS, option 1 for CAP of R-TAS from TR 38.769 is adopted with following details:
CAP duration becomes shorter with increasing value of M
FFS: exact relation between duration of CAP and M values including {2,4,6,24}


Proposal 3-2a 
For the pattern of CAP of R-TAS, only following 2 alternatives are considered [for further down-selection to one alternative]:
Alt 1: 3 chips with ON-OFF-ON pattern 
For lowest M  = 2, maximum CAP duration will be 1.5 OFDM symbols with this alternative
Alt 2: 4 chips with ON-OFF-ON-OFF pattern
For lowest M  = 2, maximum CAP duration will be 2 OFDM symbols with this alternative
Based on the minimum value of M that will be agreed in agenda 9.4.2, corresponding maximum duration of CAP will be selected from the candidate set of {1.5 OFDM symbols, 2 OFDM symbols, 3 OFDM symbols}


Proposal 5-1a 
For D2R, 1-part preamble and 1-part midamble, base sequence is generated from M-sequence, where the length of the sequence is 
Value(s) of n = 5, 3
Long preamble is generated based on n = 5
Short preamble is generated based on following n values to be down-selected to single value
Alt 1: n = 3
Alt 2: n = 4
FFS: signaling/application of the two preambles
Same base sequence is applied for both preamble and midamble
Preamble immediately precedes the PDRCH without any gap


Proposal 4-1a 
Take either Alt 1 or Al 2:

Alt 1: For indicating the end of PRDCH transmission, Manchester coding rule violation corresponding to M value is adopted
FFS: whether the Manchester coding rule violation is based on specified postamble pattern or based on device detecting such violation based on reader’s implementation
Note: If companies cannot converge on whether/what postamble is specified or not, then device detecting such violation based on reader’s implementation will be adopted
Note: If no postamble is specified, then from timing perspective, end of last chip of PRDCH is the end of R2D transmission
Alt 2: There is no consensus to specify R2D postamble




Contributions in RAN1#120bis


Appendix
Revised WID (RP-243326): RAN1 Scope & Objectives 
General Scope
The definitions provided in TR 38.848, TR 38.769, and decisions, etc. made during the Rel-19 SI in RAN WGs are taken into this WI, and the following is the exclusive general scope:
The overall objective shall be to standardize the following Ambient IoT device:
Device 1: ~1 µW peak power consumption, has energy storage, RF envelope detector receiver, initial sampling frequency offset (SFO) up to 10X ppm, neither R2D nor D2R amplification in the device. The device’s D2R transmission is backscattered on a carrier wave provided externally.
Deployment scenario 1 with Topology 1, according to D1T1-B. 
FR1 licensed spectrum in FDD, with R2D in DL spectrum and D2R and CW in UL spectrum.
Spectrum deployment in-band to NR and standalone, with A-IoT BS located indoor.
Traffic types DO-DTT, DT, for rUC1 (indoor inventory) and rUC4 (indoor command). 
Carrier wave transmission for waveform 1 only, without hopping, per the following cases in TR 38.769:
Case 1-4 for D1T1-B
Proximity determination via Solution 1 in TR 38.769 only.
Device (un)availability via Direction 1 in TR 38.769 only.

WGs begin their discussions from the decisions already made in TR 38.769, with the following refinements for the scope: 

The following objectives are set, within the General Scope:
RAN1 scope:
PRDCH and PDRCH, which are the only physical channels in R2D and D2R, respectively.
R2D and D2R signal(s)
Multiplexing/multiple access in R2D is by only TDMA, and in D2R is by only TDMA and FDMA.
R2D supports only OOK-4 modulation, one solution for CP handling. D2R backscattering supports only OOK and BPSK modulations.
R2D transmission supports only the Manchester line code in TR 38.769
D2R transmission supports:
Either the Manchester line code in TR 38.769 or no line code (one to be down-selected); and
A corresponding small frequency shift method according to the options in TR 38.769.
R2D does not support FEC. D2R supports only convolutional code with generator polynomials as per TS 36.212. Applying or not applying the FEC to D2R is specified by ensuring it is under the reader control and applies to all devices targeted by the reader.
PRDCH and PDRCH both support transmission without CRC, and with CRC as per the generator polynomials in TS 38.212 for 6-bit CRC and 16-bit CRC. Cases to use which length of CRC, or no CRC, to be decided in RAN1.
D2R supports physical layer repetition transmission. R2D does not support physical-layer repetition transmission. 
RAN2 scope:
Specify the necessary functions and procedures for an Ambient IoT compact protocol stack and lightweight signalling procedure to enable DO-DTT and DT data transmission:
A-IoT Paging, including subsequent paging for the same service. Support the options that a paging message contains one identifier, and that a paging message contains no identifier. 
Note: RAN2 aims to design a paging message format such that multiple identifiers can be contained in one paging message, for forward compatibility purposes.
A-IoT Random access, including re-access for failure handling. Contention-based and contention-free cases are supported. For the contention-based random access, only Solution 1 (3-step only) is included.
A-IoT data transmission, including data (re-)transmission for failure handling. Segmentation is supported at least in D2R.
Only MAC layer is included
RAN3 scope: 
Specify necessary architectural aspects, and signaling and procedures between A-IoT RAN and A-IoT CN to support the A-IoT functions, assuming an architecture of aggregated gNB, including:
Inventory and command operations
Device location reporting at reader ID granularity
Note: The above A-IoT functions are supported over the existing NG interface, based on architecture(s) defined by RAN3/SA2.
RAN4 scope:
Specify RF requirements for Ambient-IoT BS, device 1, and CW
RF requirements for Type 1-C Ambient-IoT BS
RF requirements for device 1
RF requirements for CW
Specify RRM core requirements for device 1, if necessary
Study and develop OTA test methodology for A-IoT device 1
Consider test methods specified in TR 38.870 as starting point. Take test system reuse, test system complexity and test time into account, when developing test methods suitable for Ambient IoT.
Develop the preliminary Measurement Uncertainty (MU) assessment for the test system
Use band n8 as an example band

Note 1: Coordination with SA2 and SA3 is expected. Updates to the WID objectives should be considered if needed.

Note 2: This WI shall target for an IoT segment well below the existing 3GPP IoT technologies, e.g. NB-IoT, eMTC, RedCap, etc. The WI shall not aim to replace existing 3GPP LPWA technologies.

SI Phase: RAN1 Agreements (relevant for R2D/D2R signals including timing acquisition and synchronization
RAN1#116 (Athens, Greece, February 26th – March 1st, 2024)
Agreement
At least the following time domain frame structure is studied for A-IoT R2D and D2R transmission.
For R2D transmission,
A R2D timing acquisition signal (e.g. R2D preamble) is included at least for timing acquisition and for indicating the start of the R2D transmission in time domain.
For D2R transmission,
A D2R timing acquisition signal (e.g. D2R preamble) is included at least for timing acquisition and for indicating the start of the D2R transmission in time domain.
FFS other necessary component(s), e.g. midamble, postamble, periodic sync signal, control fields, guard period


RAN1#116bis (Changsha, Hunan Province, China, April 15th – April 19th, 2024)

Agreement
To determine or derive the end of PRDCH transmission, study at least following options:  
Option 1: R2D postamble immediately follows the PRDCH to indicate the end of the PRDCH.       
Option 2: Based on R2D control information.

Agreement
For the reader to acquire the end of PDRCH transmission, study at least following options:  
Option 1: D2R postamble immediately follows the PDRCH
Option 2: Based on control information

Agreement
For D2R transmission, study the necessity of midamble at least for the purpose of performing timing/frequency tracking or channel estimation or interference estimation, considering at least the following: 
Modulation and Coding schemes, e.g., data modulation, line/channel coding 
Receiving methods, e.g., coherent or non-coherent
D2R transmission length/packet size
Midamble overhead
Timing/frequency accuracy
Phase accuracy

Agreement
RAN1 study the R2D transmission without midamble as the baseline if Manchester encoding is used.
FFS the necessity for the R2D transmission with midamble if PIE is used. 

Agreement
For the R2D timing acquisition signal immediately preceding the transmission of a physical channel, study a preamble with at least two parts which includes a start-indicator part and a clock-acquisition part, where the start-indicator part immediately precedes the clock-acquisition part:
Start-indicator part provides the start of the R2D transmission
FFS: Details of start-indicator part
Clock-acquisition part provides at least the chip synchronization of the subsequent physical channel transmission
FFS: Details of clock-acquisition part, e.g. structure, encoding, length, etc. 
FFS: Methods to determine chip duration of the subsequent physical channel transmission 
FFS: Other functionalities
Note: the preamble is considered not to be part of a physical channel
FFS: other part(s) of the preamble, if any 
FFS: whether the above clock acquisition is sufficient for all devices
FFS: how to make the preamble compact

Agreement
For D2R, a preamble preceding each PDRCH transmission is studied as the baseline at least for the D2R timing acquisition signal:
Preamble is not part of PDRCH
FFS: Other functionalities of the preamble

Agreement
Reference signals including at least DMRS, PTRS, CSI-RS/TRS, are not further studied for R2D.

Agreement
Reference signals including DMRS, PTRS, SRS, are not further studied for D2R
Note: This doesn’t preclude the possibility to study preamble, midamble, postamble for different purposes, e.g. channel/interference estimation and/or proximity determination


RAN1#117 (Fukuoka City, Fukuoka, Japan, May 20th – 24th, 2024)

Agreement
For the start-indicator part of the R2D time acquisition signal, study the two options below:
Option 1: ON/OFF pattern i.e. high/low voltage transmission 
Option 2: OFF pattern, i.e. low voltage transmission 

Agreement
For R2D, the clock-acquisition part of the R2D time acquisition signal is used to determine the OOK chip duration
FFS: Pattern design to support determination of chip duration


RAN1#118 (Maastricht, NL, August 19th – 23rd,  2024)

Agreement
For each D2R transmission, no separate part for start-indicator is considered for the preamble preceding the PDRCH.

Agreement
For D2R transmission, preamble preceding the PDRCH is studied also for the potential additional functionalities:
SFO estimation
CFO estimation
Channel estimation
Interference estimation
Note: this does not preclude studying the above functionalities by using a midamble and/or postamble, if supported
FFS: Other functionalities, if any

Agreement
For the start-indicator part of the R2D time acquisition signal, ON/OFF pattern i.e. high/low voltage transmission is applied
FFS: length/pattern of ON/OFF.
FFS: when TD2R_min is applicable, whether/how the start-indicator part is included in TD2R_min or not. To be discussed in 9.4.2.2


RAN1#118bis (Hefei, China, October 14th – 18th,  2024)

Agreement
The start indicator part of the R2D time acquisition signal is not included in TD2R_min.

Agreement
The TR will capture the following options, and companies are encouraged to analyze the tradeoffs among the following D2R amble(s) options:
Option 1: D2R preamble only
Option 2: D2R preamble + X midamble(s), where X 1
Option 3: D2R preamble + postamble
Option 4: D2R preamble + Y midamble(s) + postamble, where Y1
For the above options, companies are encouraged to report at least the following:
Purpose(s) of the preamble, midamble and postamble 
Whether companies assume multiple options can be supported


Agreement
For analysing the trade-offs among the D2R amble(s) options, companies can refer to the Table 3.2.4 in section 3.2.4 of R1-2408993 for information.

Agreement
For the clock-acquisition part of the R2D time acquisition signal, following is captured in the TR 38.769:
Clock-acquisition part is based on OOK without line coding and includes rising/falling edges, including at least two rising or at least two falling edges for the device to determine the OOK chip duration

Agreement
For the start-indicator part of the R2D time acquisition signal, for providing the start of the R2D transmission, following is captured in the TR 38.769:
Following options have been studied for the start-indicator part of the R2D time acquisition signal:
Option 1: ON-OFF transmission is considered based on energy/edge detection, and multiple alternatives have been studied including 
Alt 1: A single ON-OFF transmission, i.e. one high-voltage transmission followed by one low-voltage transmission, where ON and OFF may have same or different durations
Alt 2: A multi-ON-OFF transmission, where different ON and different OFF may have same or different durations and different parts may have same or different duration
Option 2: ON-OFF sequence-based design is considered which consists of a pre-defined sequence for detection of start-indicator part based on digital correlation
For both the options, it is observed that a fixed duration for the start-indicator part can be considered, regardless of the value of M used for PRDCH transmissions. 
Miss-detection ratio (MDR), false-alarm ratio (FAR) and detection complexity have been considered for the design of the R2D start indicator part by following companies
It is observed by 1 source [Huawei] that for an FAR of ~0%, the MDR of less than 1% can be achieved with Alt 2 of option 1 (considering 2 ON-OFF transmissions with different durations) and it is also observed that low-complexity and reduced power consumption can be achieved
1 source [ZTE] evaluated Alt 1 of option 1 (considering same duration for ON and OFF) and Alt 2 of option 1 (considering multiple ON-OFF transmissions with same duration) and observed that for an FAR of ~0%, the MDR of less than 1% can be achieved and Alt 1 of option 1 performs better than Alt 2 of option 1. 
1 source [CATT] observed with ON-OFF pattern, that for an FAR of ~0%, the MDR of less than 1% can be achieved with a duration of at least 1 OFDM symbol
1 source [Qualcomm] compares the performance between option 1 and option 2. It shows almost similar coverage range (SNR requirement) for target MDR of 1%. For MDR of 10%, it shows that sequence-based design provides better performance, and it is observed that during the available time, it is feasible for all devices to detect the start-indicator sequence within the power budget. It is further observed that the FAR with sequence-based design can be improved in case of interference scenarios when compared with pattern-based design. 
For both the options, it may be beneficial that the start-indicator part is distinguishable at least from other parts of the R2D transmissions

Agreement
For the clock-acquisition part of the R2D time acquisition signal for OOK chip duration determination, following options are studied:
Option 1: Duration of the clock-acquisition part is variable for different M values, i.e. the duration becomes shorter with increasing value of M
Option 2: Duration of the clock-acquisition part is constant for different M values based on repetition, i.e. repetition factor is increased with increasing value of M to keep the duration constant
FFS: Whether/what restriction on M values for the clock-acquisition part
Note: Other functionalities of clock-acquisition part is a separate discussion

Agreement
For the D2R preamble, binary signal is considered.



RAN1#119 (Orlando, US, Nov 18th – 22nd, 2024)
Agreement
Capture following observations in the TR 38.769, where CFO is assumed to be zero or negligible.
For coherent detection of PDRCH with a payload of 16 bits or 20 bits with 6-bit or 16-bit CRC, using 1/2 Manchester coding and 1/3 or 1/2 convolutional code:
Source [6, ZTE] observed that with up to 10% SFO, ~1kbps data rate, when the same amble(s) overhead is maintained, Option 3 provides comparable performance results to Option 1.
Source [7, Samsung] observed that with up to 10% SFO, ~5kbps data rate, for device 1 and with up to 1% SFO for device 2, the decoding performance with/without midamble are similar
Source [9, vivo] observed that Option 1 is sufficient to achieve 10% and 1% BLER, with no more than 8 SFO hypotheses tested at the reader side.
With up to 10% SFO, ~ 5kbps data rate, the SNR needed to achieve 10% and 1% BLER is similar (~ -2dB and 4 dB) for Option 1, Option 2 of D2R preamble+1midamble and Option 3.
With up to 1% SFO, ~ 5kbps data rate, the SNR needed to achieve 10% and 1% BLER is similar (~ -2.8dB and 3.3dB) for Option 1, Option 2 of D2R preamble+1 midamble and Option 3.
Source [20, OPPO] observed that with perfect SFO estimation, 1kbps data rate and OOK modulation, there is no noticeable performance gain from using midamble(s) and/or postamble for PDRCH transmission, regardless of the payload size.
Source [30, Qualcomm] observed that the required SFO estimation accuracy to achieve 1% and 10% BLER is <5*10^3 ppm
With up to 10% SFO, achieving the required accuracy necessitates more than 20 SFO hypotheses at the reader side for Option 1 and 10 SFO hypotheses are sufficient for Option 3 of D2R preamble + postamble. But for Option 3 reader has to store the received samples and wait for the postamble that is after the end of PDRCH for any of SFO/channel/interference estimation, demodulation, decoding, etc. 
With up to 1% SFO, 4 SFO hypotheses are sufficient for Option 1 to achieve the required accuracy.

For coherent detection of PDRCH with a payload of 96bits with 16-bit CRC (or 6-bit CRC [14, Xiaomi]), using 1/2 Manchester coding and 1/3 or 1/2 convolutional code,
Sources [3, Huawei], [5, CMCC] and [14, xiaomi] observed that Option 1 cannot achieve 10% BLER.
Sources [6, ZTE], [7, Samsung], [9, vivo], [20, OPPO] and [30, QC] observed that Option 1 can achieve 10% BLER.
Sources [3, Huawei], [5, CMCC], [6, ZTE], [7, Samsung], [9, vivo], [14, xiaomi], [16, China Telecom] observed that adding additional amble improves the performance. 
Source [3, Huawei] observed that with up to 10% SFO, 
Option 2 of D2R preamble+ 1 midamble achieves 10% BLER at SNR around -3dB, but cannot achieve 1% BLER.
Option 3 of D2R preamble+ postamble achieves 10% BLER at SNR around -4dB, and can achieve 1% BLER at SNR around 4dB.
Source [5, CMCC] observed that with up to 10% SFO, Option 3 allows reader to precisely search and detect the SFO with 0.03% residual SFO at -3dB SNR TDL-A channel, achieving 10% BLER -2.44dB SNR for ~1 kbps data rate and -2.17 dB for ~6 kbps data rate. Source [5, CMCC] further observed that when the reader adopts same number of SFO hypothesis based on preamble, with 1% SFO, Option 3 can achieve 10% BLER at -4.27 dB SNR for ~1 kbps and at -4.29 dB SNR for ~6 kbps, which provides 1~2 dB performance gain when compared to 10% SFO. 
Source [14, xiaomi] observed that with up to 10% SFO, Option 2 of D2R preamble + 1 midamble, option 3, and option 4 of D2R preamble + 1 midamble+postamble achieve basically the same performance, the SNR for 10% BLER is 5dB for 1.25 kbps data rate.  
Source [6, ZTE] observed that with up to 10% SFO, ~1kbps data rate, and the same amble(s) overhead, Option 3 can provide 1~2 dB, 5dB performance gain for 10% BLER, 1% BLER, respectively, compared to Option 1. Additionally, Option 3 can provide ~1dB, 2dB performance gain for 10% BLER, 1% BLER, respectively, compared to Option 2 of D2R preamble +1 midamble.
Source [7, Samsung] observed that for ~5kbps data rate, compared to option 1, 
For device 1 with up to 10% SFO, Option 2 of D2R preamble + 1 midamble provides ~0.5 dB SNR gain at 10% BLER with TDL-A channel and ~0.9 dB SNR gain with TDL-D channel.
For device 2 with up to 1% SFO, Option 2 of D2R pramble + 1midamble provides ~1 dB SNR gain at 10% BLER with TDL-A channel and ~1.4 dB SNR gain with TDL-D channel.
Source [9, vivo] observed that, 
With up to 10% SFO, ~5.5kbps data rate, Option 1 achieves 10% and 1% BLER at SNR ~0.7dB and 10dB, respectively; Additionally, maintaining the same amble overhead, Option 2 (D2R preamble + 1 midamble) and Option 3 demonstrate similar performance, achieving 10% and 1% BLER at SNR around -1.7dB and 5.2dB, respectively.
With up to 1% SFO, ~5.5kbps data rate, Option 1 achieves 10% and 1% BLER at SNR ~ -1.3dB and 11dB, respectively. Additionally, with the same amble overhead, the SNR difference between Option 2 (D2R preamble+1midamble) and Option 3 for 10% and 1% BLER is less than 1dB, with SNRs ~ -3.1dB to -2.5dB for 10% BLER and ~3.6dB to 4.5dB for 1% BLER.
Source [16, China Telecom] observed that with up to 10% SFO, ~7.5kbps data rate, there is ~6~7dB performance gap at 10% BLER and ~10.5~11.5dB performance gap at 1% BLER between option 2 of D2R preamble+111 midambles and option 1. Note that Source [16, China Telecom] does not use any convolutional code.
Source [20, OPPO] observed that with perfect SFO estimation, 1kbps data rate and OOK modulation, there is no noticeable performance gain from using midamble(s) and/or postamble for PDRCH transmission, regardless of the payload size.
Source [30, Qualcomm] observed that the required SFO estimation accuracy to achieve 1% and 10% BLER is <10^3 ppm. To achieve the required accuracy,
For Option 1, more than 50 SFO hypotheses at reader side are necessary for device with up to 10% SFO and 6 SFO hypotheses are sufficient at reader side for device with up to 1% SFO. 
For Option 3, 10 SFO hypotheses are sufficient for device with up to 10% SFO, but reader has to store the received samples and wait for the postamble that is after the end of PDRCH for any of SFO/channel/interference estimation, demodulation, decoding, etc.

For coherent detection of PDRCH with a payload of 400bits with 16-bit CRC, using 1/2 Manchester coding and 1/3 or 1/2 convolutional code, 
For option 1 of D2R preamble only, 
Sources [3, Huawei], [5, CMCC], [6, ZTE], [8, Spreadtrum], [9, vivo], [14, xiaomi] observed that with up to 10% SFO, 10% BLER cannot be achieved. 
Source [20, OPPO] observed that with perfect SFO estimation, 1kbps data rate and OOK modulation, there is no noticeable performance gain from using midamble(s) and/or postamble for PDRCH transmission.
For other amble options, 
Source [3, Huawei] observed that
With accurate SFO estimation, Option 2 of D2R preamble + 4 midambles can achieve 10% BLER at SNR ~ 2.7dB but cannot achieve 1% BLER.
With up to 10% SFO, Option 3 cannot achieve 10% BLER.
With up to 10% SFO, Option 4 of D2R preamble+2 midambles+postamble achieves 10% BLER at SNR of ~0.25dB; But it cannot achieve 1% BLER. Option 4 of D2R preamble+3 or 4 midambles+postamble, achieves a 10% BLER at an SNR of around -0.2 dB, and achieves 1% BLER at SNR around 9dB or 8dB, respectively.
Source [5, CMCC] observed that with up to 10% SFO, Option 4 of D2R preamble combined with 1 to 4 midambles + postamble, achieves 10% BLER at SNR of 2.5 dB, 1 dB, 0.8 dB, or 0.5 dB, respectively, for a data rate of around 1 kbps.
Source [6, ZTE] observed that with up to 10% SFO, 
Option 3 can provide ~5.5 dB performance gain compared to option 2 of D2R preamble+1midamble for 10% BLER, with the same amble(s) overhead for ~1kbps data rate.
Option 2 of D2R preamble+1midamble cannot achieve 1% BLER for ~1kbps data rate.
Option 4 of the D2R preamble+1 or 2 midamble(s)+postamble, has similar performance, it can achieves a 10% BLER at SNR of -1dB and achieves a 1% BLER at SNR of 6dB and 5dB respectively for ~1kbps data rate.  
Source [8, Spreadtrum] observed that with up to 10% SFO, 
Option 3 of D2R preamble+ postamble cannot achieve 10% BLER for ~7kpbs.  
Option 4 of D2R preamble + 1 midamble + postamble can achieve 10% BLER and 1% BLER at SNR around -6dB and 0 dB, respectively for ~7kpbs data rate.
Source [9, vivo] observed that 
With up to 10% SFO, maintaining the same amble overhead, both Option 2 of D2R preamble+5 midambles and Option 4 of D2R preamble+4 midambles+1 postamble demonstrate similar performance, achieving 10% BLER at SNR ~0.2dB. While for 1% BLER, the SNR for Option 2 and Option 4 is ~ 9.2dB and 12.8dB, respectively for ~5.5kpbs data rate.   
With up to 1% SFO, maintaining the same amble overhead, both Option 2 of D2R preamble+5 midambles and Option 4 of D2R preamble+4 midambles+1postamble demonstrate similar performance, achieving 10% BLER at SNR around -1.2dB. While for 1% BLER, the SNR for Option 2 and Option 4 is ~7.8dB and 9.1dB, respectively for ~5.5kpbs data rate.   
Source [14, xiaomi] observed that with up to 10% SFO, Option 2 of D2R preamble+3 midambles and Option 4 of D2R preamble+3 midambles+postamble can achieve 10% BLER when the SNR is within the range of [15, 25] dB for 1.25 kbps data rate. 
Source [30, Qualcomm] observed that the required SFO estimation accuracy to achieve 1% and 10% BLER is much smaller than 10^3 ppm. To achieve the required accuracy,
For Option 2 of D2R preamble+X midamble(s) where midamble inserted per every certain number of PDRCH bits (e.g., 192 bits),
For SFO estimation using each amble for the subsequent PDRCH bits (e.g., 192 bits), with up to 10% SFO, more than 50 SFO hypotheses are necessary at the reader side and with up to 1% SFO, 6 SFO hypotheses are sufficient at the reader side.
For SFO estimation based on the time gap between preamble and midamble, with up to 10% SFO, 10 SFO hypotheses are used, but reader has to store the received samples and wait for the midamble to start SFO/channel/interference estimation, demodulation, decoding, etc.
For Option 3 of D2R preamble+postamble, SFO estimation is based on the time gap between preamble and postamble, with up to 10% device SFO, 10 SFO hypotheses are used for reader, but reader has to store the received samples and wait for the postamble that is after the end of PDRCH for any of SFO/channel/interference estimation, demodulation, decoding, etc

For the synchronization and timing tracking of D2R transmission, 
Source [5, CMCC] report that with up to 10% SFO, option 1 is not sufficient for D2R reception since the residual SFO at reader side is larger than 1%. While with option 3, the reader can precisely search and detect the SFO with a residual SFO of 0.03% at -3dB SNR TDL-A channel.
Source [14, xiaomi] report that 
For packet size of 96bits, when the SNR is increased from -4dB to 20dB, the ratio of device residual SFO over 100ppm decreases to 6% for Option 2, 3 and 4, but remains at 95% for Option 1.
For packet size of 400bits, when the SNR is increased from -4dB to 20dB, the ratio of device residual SFO larger than 10ppm decreases to 5% for Option 2, 3, and 4, but is still 99.6% for Option 1.
Sources [9, vivo], [15, CATT] report that SFO estimation based on D2R preamble can achieve accurate estimation without additional ambles (midamble or postamble). 
Source [9, vivo][7 Samsung] observed that for non-coherent detection of PDRCH, the number of SFO hypotheses and the SNR needed for 10% and 1% BLER cannot significantly be reduced for option 2, 3 and 4 compared to the option 1. Moreover, the additional ambles i.e., midamble(s) and/or postamble introduces additional overhead and prevents pipelined processing of the reception.     
Source [15, CATT] observed that 
The coarse estimation of SFO based on the D2R preamble indicates that the SFO estimation error is less than 1% with a probability of 99.3%, and less than 0.1% with a probability of 49.9%.
The fine estimation of SFO based on the D2R preamble shows that the SFO estimation error is less than 1% with a probability of 99.5%, and less than 0.1% with a probability of 90.8%.
Reader/gNB can achieve a probability of not less than 99.5% for SFO estimation error below 1%, and 90.8% for SFO estimation error below 0.1% by receiving D2R preamble signals.
Source [30, Qualcomm] report that for D2R with coherent demodulation at reader, the reader needs to estimate the device clock frequency with the accuracy of 0.5% (5 * 10^3 ppm) or lower for a short message (e.g., 72 bits after CRC/coding) and 0.1% (10^3 ppm) or lower for a long message (e.g., 224 bits after CRC/coding). The source further reports that design of D2R amble(s) (e.g., overhead) and the correspondingly required number of SFO hypothesis for the estimation depend on the sampling clock accuracy that the device uses for D2R. 
Note: in the observations above where coherent detection is used, sources that evaluated option 3 and option 4 assumed that the postamble is used at least for time/frequency tracking and for channel estimation.


Agreement
For the CFO calibration signal, which is required only for device 2b to reduce the frequency offset range and the guard-bandwidth of D2R transmission, the following observations are captured in TR 38.769:
Source [3, Huawei] report that a single-tone RF signal is used as the CFO calibration signal, it is not a part of time acquisition signal and can be transmitted as an optional R2D signal after the PRDCH transmission. 
Sources [2, Ericsson], [19, Panasonic] and [20, OPPO] report that additional synchronization signal is needed. 
[OPPO] state the R2D timing acquisition signal may not be sufficient or may not be usable for CFO calibration since a reference frequency is needed when separate LOs are used for Tx and Rx in device 2b.
Sources [7, Samsung], [9, vivo], [30, Qualcomm], [36, Apple] report that additional synchronization signal is needed if the synchronization for carrier frequency using R2D signal/channel does not provide required functionalities for device 2b.
Source [5, CMCC][31, MTK] report that it may not be possible to achieve enough frequency accuracy (0.01 ppm) even after CFO calibration based on R2D time acquisition signals for coherent detection at reader especially when the D2R data rate is low.

Agreement
For device 2b, a signal for CFO calibration should be provided to synchronize / calibrate the device clock for LO for carrier frequency (Clock purpose #5) to achieve the accuracy after clock sync / calibration at device side captured in Table 5.2.3-1.
Frequency calibration at device 2b is beneficial at least to reduce the guard-bandwidth of D2R transmission.

Agreement
Adopt the updates documented in R1-2410653 for section 6.2 of the TR38.769. 

Agreement
Adopt following update to the TP agreed on Monday

Capture following observations in the TR 38.769, where CFO is assumed to be zero or negligible.
[omit unchanged part]
For coherent detection of PDRCH with a payload of 96bits with 16-bit CRC (or 6-bit CRC [14, Xiaomi]), using 1/2 Manchester coding and 1/3 or 1/2 convolutional code,
Sources [3, Huawei], [5, CMCC] and [14, xiaomi] observed that Option 1 cannot achieve 10% BLER.
Sources [6, ZTE], [7, Samsung], [9, vivo], [20, OPPO] and [30, QC] observed that Option 1 can achieve 10% BLER.
Sources [3, Huawei], [5, CMCC], [6, ZTE], [7, Samsung], [9, vivo], [14, xiaomi], [16, China Telecom] observed that adding additional amble improves the performance. 
Source [3, Huawei] observed that with up to 10% SFO, 
Option 2 of D2R preamble+ 1 midamble achieves 10% BLER at SNR around -3dB, but cannot achieve 1% BLER.
Option 3 of D2R preamble+ postamble achieves 10% BLER at SNR around -4dB, and can achieve 1% BLER at SNR around 4dB.
Source [5, CMCC] observed that with up to 10% SFO, Option 3 allows reader to precisely search and detect the SFO with 0.03% residual SFO at -3dB SNR TDL-A channel, achieving 10% BLER -2.44dB SNR for ~1 kbps data rate and -2.17 dB for ~6 kbps data rate. Source [5, CMCC] further observed that when the reader adopts same number of SFO hypothesis based on preamble, with 1% SFO, Option 3 can achieve 10% BLER at -4.27 dB SNR for ~1 kbps and at -4.29 dB SNR for ~6 kbps, which provides 1~2 dB performance gain when compared to 10% SFO. 
Source [14, xiaomi] observed that with up to 10% SFO, Option 2 of D2R preamble + 1 midamble, option 3, and option 4 of D2R preamble + 1 midamble+postamble achieve basically the same performance, the SNR for 10% BLER is 5dB for 1.25 kbps data rate.  
Source [6, ZTE] observed that with up to 10% SFO, ~1kbps data rate, and the same amble(s) overhead, Option 3 can provide 1~2 dB, 5dB performance gain for 10% BLER, 1% BLER, respectively, compared to Option 1. Additionally, Option 3 can provide ~1dB, 2dB performance gain for 10% BLER, 1% BLER, respectively, compared to Option 2 of D2R preamble +1 midamble.
Source [7, Samsung] observed that for ~5kbps data rate, compared to option 1, 
For device 1 with up to 10% SFO, Option 2 of D2R preamble + 1 midamble provides ~0.5 dB SNR gain at 10% BLER with TDL-A channel and ~0.9 dB SNR gain with TDL-D channel.
For device 2 with up to 1% SFO, Option 2 of D2R pramble + 1midamble provides ~1 dB SNR gain at 10% BLER with TDL-A channel and ~1.4 dB SNR gain with TDL-D channel.
Moreover, the additional ambles i.e., midamble(s) and/or postamble introduces additional overhead and postamble may prevent pipelined processing of the reception. 
Source [9, vivo] observed that, 
With up to 10% SFO, ~5.5kbps data rate, Option 1 achieves 10% and 1% BLER at SNR ~0.7dB and 10dB, respectively; Additionally, maintaining the same amble overhead, Option 2 (D2R preamble + 1 midamble) and Option 3 demonstrate similar performance, achieving 10% and 1% BLER at SNR around -1.7dB and 5.2dB, respectively.
With up to 1% SFO, ~5.5kbps data rate, Option 1 achieves 10% and 1% BLER at SNR ~ -1.3dB and 11dB, respectively. Additionally, with the same amble overhead, the SNR difference between Option 2 (D2R preamble+1midamble) and Option 3 for 10% and 1% BLER is less than 1dB, with SNRs ~ -3.1dB to -2.5dB for 10% BLER and ~3.6dB to 4.5dB for 1% BLER.
Source [16, China Telecom] observed that with up to 10% SFO, ~7.5kbps data rate, there is ~6~7dB performance gap at 10% BLER and ~10.5~11.5dB performance gap at 1% BLER between option 2 of D2R preamble+111 midambles and option 1. Note that Source [16, China Telecom] does not use any convolutional code.

Source [20, OPPO] observed that with perfect SFO estimation, 1kbps data rate and OOK modulation, there is no noticeable performance gain from using midamble(s) and/or postamble for PDRCH transmission, regardless of the payload size.
Source [30, Qualcomm] observed that the required SFO estimation accuracy to achieve 1% and 10% BLER is <10^3 ppm. To achieve the required accuracy,
For Option 1, more than 50 SFO hypotheses at reader side are necessary for device with up to 10% SFO and 6 SFO hypotheses are sufficient at reader side for device with up to 1% SFO. 
For Option 3, 10 SFO hypotheses are sufficient for device with up to 10% SFO, but reader has to store the received samples and wait for the postamble that is after the end of PDRCH for any of SFO/channel/interference estimation, demodulation, decoding, etc.

For coherent detection of PDRCH with a payload of 400bits with 16-bit CRC, using 1/2 Manchester coding and 1/3 or 1/2 convolutional code, 
For option 1 of D2R preamble only, 
Sources [3, Huawei], [5, CMCC], [6, ZTE], [8, Spreadtrum], [9, vivo], [14, xiaomi] observed that with up to 10% SFO, 10% BLER cannot be achieved. 
Source [20, OPPO] observed that with perfect SFO estimation, 1kbps data rate and OOK modulation, there is no noticeable performance gain from using midamble(s) and/or postamble for PDRCH transmission.
For other amble options, 
Source [3, Huawei] observed that
With accurate SFO estimation, Option 2 of D2R preamble + 4 midambles can achieve 10% BLER at SNR ~ 2.7dB but cannot achieve 1% BLER.
With up to 10% SFO, Option 3 cannot achieve 10% BLER.
With up to 10% SFO, Option 4 of D2R preamble+2 midambles+postamble achieves 10% BLER at SNR of ~0.25dB; But it cannot achieve 1% BLER. Option 4 of D2R preamble+3 or 4 midambles+postamble, achieves a 10% BLER at an SNR of around -0.2 dB, and achieves 1% BLER at SNR around 9dB or 8dB, respectively.
Source [5, CMCC] observed that with up to 10% SFO, Option 4 of D2R preamble combined with 1 to 4 midambles + postamble, achieves 10% BLER at SNR of 2.5 dB, 1 dB, 0.8 dB, or 0.5 dB, respectively, for a data rate of around 1 kbps.
Source [6, ZTE] observed that with up to 10% SFO, 
Option 3 can provide ~5.5 dB performance gain compared to option 2 of D2R preamble+1midamble for 10% BLER, with the same amble(s) overhead for ~1kbps data rate.
Option 2 of D2R preamble+1midamble cannot achieve 1% BLER for ~1kbps data rate.
Option 4 of the D2R preamble+1 or 2 midamble(s)+postamble, has similar performance, it can achieve a 10% BLER at SNR of -1dB and achieves a 1% BLER at SNR of 6dB and 5dB respectively for ~1kbps data rate.  
Source [8, Spreadtrum] observed that with up to 10% SFO, 
Option 3 of D2R preamble+ postamble cannot achieve 10% BLER for ~7kpbs.  
Option 4 of D2R preamble + 1 midamble + postamble can achieve 10% BLER and 1% BLER at SNR around -6dB and 0 dB, respectively for ~7kpbs data rate.
Source [9, vivo] observed that 
With up to 10% SFO, maintaining the same amble overhead, both Option 2 of D2R preamble+5 midambles and Option 4 of D2R preamble+4 midambles+1 postamble demonstrate similar performance, achieving 10% BLER at SNR ~0.2dB. While for 1% BLER, the SNR for Option 2 and Option 4 is ~ 9.2dB and 12.8dB, respectively for ~5.5kpbs data rate.   
With up to 1% SFO, maintaining the same amble overhead, both Option 2 of D2R preamble+5 midambles and Option 4 of D2R preamble+4 midambles+1postamble demonstrate similar performance, achieving 10% BLER at SNR around -1.2dB. While for 1% BLER, the SNR for Option 2 and Option 4 is ~7.8dB and 9.1dB, respectively for ~5.5kpbs data rate.   
Source [14, xiaomi] observed that with up to 10% SFO, Option 2 of D2R preamble+3 midambles and Option 4 of D2R preamble+3 midambles+postamble can achieve 10% BLER when the SNR is within the range of [15, 25] dB for 1.25 kbps data rate. 
Source [30, Qualcomm] observed that the required SFO estimation accuracy to achieve 1% and 10% BLER is much smaller than 10^3 ppm. To achieve the required accuracy,
For Option 2 of D2R preamble+X midamble(s) where midamble inserted per every certain number of PDRCH bits (e.g., 192 bits),
For SFO estimation using each amble for the subsequent PDRCH bits (e.g., 192 bits), with up to 10% SFO, more than 50 SFO hypotheses are necessary at the reader side and with up to 1% SFO, 6 SFO hypotheses are sufficient at the reader side.
For SFO estimation based on the time gap between preamble and midamble, with up to 10% SFO, 10 SFO hypotheses are used, but reader has to store the received samples and wait for the midamble to start SFO/channel/interference estimation, demodulation, decoding, etc.
For Option 3 of D2R preamble+postamble, SFO estimation is based on the time gap between preamble and postamble, with up to 10% device SFO, 10 SFO hypotheses are used for reader, but reader has to store the received samples and wait for the postamble that is after the end of PDRCH for any of SFO/channel/interference estimation, demodulation, decoding, etc
Source [7, Samsung] observes that the additional ambles i.e., midamble(s) and/or postamble introduces additional overhead and postamble may prevent pipelined processing of the reception.

For the synchronization and timing tracking of D2R transmission, 
Source [5, CMCC] report that with up to 10% SFO, option 1 is not sufficient for D2R reception since the residual SFO at reader side is larger than 1%. While with option 3, the reader can precisely search and detect the SFO with a residual SFO of 0.03% at -3dB SNR TDL-A channel.
Source [14, xiaomi] report that 
For packet size of 96bits, when the SNR is increased from -4dB to 20dB, the ratio of device residual SFO over 100ppm decreases to 6% for Option 2, 3 and 4, but remains at 95% for Option 1.
For packet size of 400bits, when the SNR is increased from -4dB to 20dB, the ratio of device residual SFO larger than 10ppm decreases to 5% for Option 2, 3, and 4, but is still 99.6% for Option 1.
Sources [9, vivo], [15, CATT] report that SFO estimation based on D2R preamble can achieve accurate estimation without additional ambles (midamble or postamble). 
Source [9, vivo][7 Samsung] observed that for non-coherent detection of PDRCH, the number of SFO hypotheses and the SNR needed for 10% and 1% BLER cannot significantly be reduced for option 2, 3 and 4 compared to the option 1. Moreover, the additional ambles i.e., midamble(s) and/or postamble introduces additional overhead and postamble may prevents pipelined processing of the reception.     
Source [15, CATT] observed that 
The coarse estimation of SFO based on the D2R preamble indicates that the SFO estimation error is less than 1% with a probability of 99.3%, and less than 0.1% with a probability of 49.9%.
The fine estimation of SFO based on the D2R preamble shows that the SFO estimation error is less than 1% with a probability of 99.5%, and less than 0.1% with a probability of 90.8%.
Reader/gNB can achieve a probability of not less than 99.5% for SFO estimation error below 1%, and 90.8% for SFO estimation error below 0.1% by receiving D2R preamble signals.
Source [30, Qualcomm] report that for D2R with coherent demodulation at reader, the reader needs to estimate the device clock frequency with the accuracy of 0.5% (5 * 10^3 ppm) or lower for a short message (e.g., 72 bits after CRC/coding) and 0.1% (10^3 ppm) or lower for a long message (e.g., 224 bits after CRC/coding). The source further reports that design of D2R amble(s) (e.g., overhead) and the correspondingly required number of SFO hypothesis for the estimation depend on the sampling clock accuracy that the device uses for D2R. 
Source [37, MediaTek] reports that transmitting 96-bit packet size with 16-bit CRC requires residue SFO after reader compensation to be 1000 ppm, and transmitting 1000-bit packet size with 16-bit CRC requires residue SFO after reader compensation to be 100 ppm.
Note: in the observations above where coherent detection is used, sources that evaluated option 3 and option 4 assumed that the postamble is used at least for time/frequency tracking and for channel estimation.

Agreement
Following observations on R2D clock-acquisition part are captured in TR 38.769:
On impact/restriction of M values for the clock-acquisition part
9 sources [TCL, Nokia, Huawei, CMCC, ZTE, Apple, CATT, Mediatek, Qualcomm] provided observations on the impact/restriction of M values for the clock-acquisition part design requirements:
1 source [Nokia] observed that increasing value of M, while retaining the same transmission duration, improves the auto-/cross- correlation properties of the sequence due to increase in sequence length and use of   provides better timing estimation accuracy even in the presence of SFO as the sequence length spans only over a shorter duration. 
2 sources [TCL, Huawei] observed for option 1 of the clock-acquisition part design that no restriction is required to be placed on the M values. Furthermore, 1 source [Huawei] observed that the same 2 ON-OFF voltage (with the same duration) satisfies the FDR performance metric of less than 1% for different M values, e.g., M = 2, 6 and 24, where FDR is the False detection ratio (FDR), i.e. incorrectly calculating M, is the performance metric.
1 source [CMCC] observed that pattern of the clock-acquisition part is related to M chips per OFDM symbol and when M is small, the clock-acquisition part may cross multiple OFDM symbols, and the CP insertion may degrade the timing acquisition performance.
1 source [ZTE] observed that with option 2, the duration of the clock-acquisition part remains consistent across all M values, at least three OFDM symbols maybe required for clock-acquisition part and it maybe not as efficient as option1
1 source [Apple] observed that among the two options studied for the clock-acquisition part, option 2 provides increased robustness, especially in case of large value of M, when compared to option 1 and potentially increase the detection performance of the clock-acquisition part. 
1 source [CATT] observed that if the chip duration is variable based on the M value used for OOK-4 waveform, the detection performance would be limited by the received SINR of the CAP with clear transition of the rising and falling edges.
1 source [Qualcomm] observed that the option 1 with M>1 has shorter duration of clock acquisition part than M=1 and worse timing acquisition accuracy. At least part of PRDCH following the clock acquisition part may need to be used to improve the timing acquisition. Furthermore, the larger M (e.g., M>4) with small chip duration is more sensitive to the SFO accuracy and the restriction of M for the clock acquisition part may be needed.
1 source [Mediatek] further observed that different M values may impact the chip accuracy obtained by the clock acquisition part.

On impact of CP insertion/handling on the clock-acquisition part
10 sources [TCL, CMCC, ZTE, Samsung, Vivo, CATT, NTT Docomo, Qualcomm, Mediatek, Spreadtrum] observed that the CP insertion/handling may impact the design requirements of the clock-acquisition part:
1 source [CMCC] further observed that when the clock-acquisition part occupies more than one OFDM symbol, ON-OFF state transition around CP can avoid the error rising or falling edges due to the CP insertion.
1 source [ZTE] further observed that to mitigate the impact of the CP in the clock-acquisition part for large M values, it can reuse the CP handling method for PRDCH 
1 source [Samsung] further observed that CP insertion/handling on the clock-acquisition part can cause false rising/falling transition and, therefore, the clock acquisition part should be designed such that it does not incur a false rising or falling edges due to CP insertion when CP-OFDM is used for OOK signal generation.
1 source [vivo] further observed that CP insertion/handling on the clock acquisition part will impact the chip duration estimation accuracy. It is further observed that for CP handling, device may not be able to count the clock and estimate OFDM symbol duration accurately until the clock acquisition part if the start indicator only includes a single ON-OFF transmission. 
1 source [CATT] further observed that the SER will be degraded due to uneven chip interval when the CP is inserted within an OFDM symbol, where SER refers to the number of samples which is mismatched for comparing to the total number of samples in a chip.
1 source [NTT Docomo] further observed if CP insertion would cause false rising/falling edges, accuracy of timing acquisition may be impacted.   
1 source [Mediatek] further observed that the issues of chip extension, false raising/falling transition, and additional raising/falling transition caused by CP insertion/handling considering different M values will impact the chip accuracy obtained by the clock acquisition part.
1 source [Spreadtrum] further observed that the design of clock acquisition part should consider that CP insertion does not cause a false rising or falling edges and does not cause different length of multiple high / low voltages within the clock acquisition part when the clock acquisition spans multiple OFDM symbols.
1 source [Huawei] observed CP insertion/handling may not impact the design requirements of the clock-acquisition part

Agreement
For the D2R preamble design, following aspects have been studied and can be captured in the TR 38.769:
Autocorrelation Property
10 sources [Nokia, Huawei, CMCC, Xiaomi, CATT, Oppo, Ericsson, NTT Docomo, Qualcomm, ZTE] observed that the signal should have good autocorrelation properties for accurate peak detection based on the signal correlation at the reader 
Cross-correlation Property
7 sources [Nokia, CMCC, Oppo, Ericsson, Qualcomm, ZTE, CATT] observed that the signal should have good cross-correlation properties if multiple D2R preamble sequences are considered (e.g. for multiple access schemes (if supported) for D2R transmissions). 
Line coding
1 source [Nokia] observed that line coding may impact the autocorrelation property of the sequence. 
1 source [Huawei] observed that for D2R preamble, to apply backscattering, line coding can help improve the detection performance based on shifting the D2R signal’s frequency location away from the carrier wave
Sequence Types (not limited to below types only)
M-sequence
3 sources [Nokia, Vivo, Xiaomi] observed that m-sequence can be considered for D2R preamble mainly owing to good correlation properties.  
Golay sequence
4 sources [CMCC, Vivo, Xiaomi, Samsung] observed that Golay sequence can be considered for D2R preamble mainly owing to good correlation properties and availability of large number of distinct sequences and complementary pairs.  
Walsh sequence
1 source [Oppo] observed that Walsh sequence can be considered as a candidate for D2R preamble thanks to its good auto/cross-correlation property and flexible length
General Observations
1 source [Huawei] observed can achieve 0.97% residual SFO with 98% probability under -2.5dB SNR and 0.1% MDR with [-1/8, 1/8] chip timing error with 99.05% probability under -2.5dB SNR with D2R preamble including 2-parts with clock-like sampling frequency signal and timing-acquisition signal, having 32-length ‘1’ sequence (encoded to 64-chip Manchester code) and 32-length sequence (encoded to 64-chip Manchester code), respectively.
4 sources [TCL, CMCC, ZTE, Vivo] observed that for D2R preamble with binary signal, the timing synchronization performance is highly related to the sequence length of the preamble. Furthermore, 1 source [CMCC] observed that to achieve a BLER performance at 10%, the timing synchronization error should be less than 10%. Furthermore, 1 source [ZTE] observed that the channel estimation performance is also highly related to sequence length. 1 source [ZTE] observed that using a 32 bits preamble provides ~8 dB, ~5 dB performance gain than using 8 bits, 16 bits preamble, respectively. And using a 64 bits preamble provides ~2.5dB performance gain than using a 32 bits preamble.
1 source [Ericsson] observed that for D2R preamble with binary signal, normalized SFO estimation error of less than 10% can be achieved with a training sequence length 64 or longer. The simulated D2R preamble consisting of a Golay complementary pair can tolerate SFO up to 1% (AWGN) with up to 1 dB loss in performance for a sufficiently long preamble sequence length (32 or greater).

Agreement
For determining the end of PRDCH at the device, following two options are studied and captured in the TR 38.769:
Option 1: TBS information (via implicit/explicit L1 R2D control information)
Option 2: Postamble (at the end of PRDCH) 
14 sources [Nokia, Huawei, ZTE, CMCC, Samsung, Ericsson, Oppo, LGE, Qualcomm, Spreadtrum, Mediatek, Cewit, Ericsson, vivo] provided following observations on the above two options for determining the end of PRDCH:
3 sources [Nokia, Huawei, ZTE] observed that option 2 provide two benefits, namely, the variable payload length and to provide timing acquisition before the subsequent transmission of either PDRCH or PRDCH, thus improving the detectability at both reader and the device, respectively. Furthermore, 1 source [Huawei] observed that R2D postamble indicates the TBS with high efficiency for small packets by avoiding a large padding overhead, unlike option 1, which may require devices to perform blind detection of different PRDCH formats (if supported) and the overhead caused by the inclusion of a R2D postamble does not exceed 20% for even the smallest of message sizes and may be less than the signaling overhead caused by using a dedicated TBS indicator
1 source [CMCC] observed for option 2, that for small payload size with only a few bits, the presence of long postamble generates large resource overhead, while for large payload size with more bits, the resource overhead of postamble is smaller.
1 source [vivo] observed for option 2, that for small payload size with only a few bits, the presence of long postamble generates large resource overhead.
1 source [Samsung] observed option 2 is not strictly required, however, given the possible clock drift at a device, it may be still beneficial to also attach postamble at least for the determination of the end of PRDCH at a device. 
3 sources [Oppo, Spreadtrum, CEWiT] observed that with option 2, the false detection may be higher for shorter postamble. Source [OPPO[ observed that in contrast to option 2, it is more reliable and efficient to indicate TBS with control information in option 1
2 sources [LGE, vivo] observed that if a message type or a command ID is included in L1 control information and implicitly indicates a known size of a fixed TB, then there is no need for either option 1 or option 2
2 sources [Qualcomm, vivo] observed that option 1 has the advantages of avoiding blind detection of postamble and providing the power saving for non-target devices to skip the R2D detection.
1 source [MediaTek] observed that option 1 is feasible for the device to avoid the unnecessary reception of a TB with a specific size and thus enable power saving, e.g., when the TB has a size exceeding the allowance of the device remaining power.
1 source [Ericsson] observed option 2 is not strictly required if the end of PRDCH can be explicitly indicated by R2D control information, and it is subject to the miss-detection rate. It may be beneficial if a PRDCH postamble can serve as an additional timing acquisition signal prior to a PDRCH transmission.

Agreement
For D2R scheduling, midamble (if supported) related information can be explicitly/implicitly indicated via corresponding PRDCH.

Agreement
Following observations on R2D clock-acquisition part are additionally captured in TR 38.769:
On purpose of SFO estimation/correction based on the clock-acquisition part
3 sources [Nokia, CATT, Qualcomm] provided observations on the applicability of clock-acquisition part for frequency synchronization:
1 source [Nokia] observed that the length of preamble sequence may need to consider also the robustness against SFO
1 source [CATT] observed that device 2a/2b may require higher synchronization accuracy for signal transmission or backscattering and therefore, the design of CAP may be required to accommodate the requirement of additional frequency synchronization and clock calibration for Device 2a/2b. 
1 source [Qualcomm] further observed for Option 1, as the CAP duration with high M is decreased, only CAP may not be sufficient for SFO correction and for Option 2, as the CAP duration is fixed and independent from M, the CAP with long enough duration can support SFO correction.
On purpose of CFO estimation/correction based on the clock-acquisition part
2 sources [Ericsson, Qualcomm] provided observations on the applicability of clock-acquisition part for frequency synchronization:
1 source [Ericsson] observes that the clock-acquisition part can be utilized to solve the frequency synchronization problem without impacting the time-domain sequence, for example by transmitting in some frequency resources and it can be a harmonized solution for both chip duration indication and device frequency synchronization. However, it is further observed that if the time interval between an R2D transmission and the corresponding D2R transmission and if the device loses the timing obtained from the R2D timing acquisition signal due to timing drift at the time for the D2R transmission, then an additional synchronization signal is needed
1 source [Qualcomm] further observed for Option 1, as the CAP duration with high M is decreased, only CAP may not be sufficient for CFO correction and for Option 2, as the CAP duration is fixed and independent from M, the CAP with long enough duration can support CFO correction.


WI Phase: RAN1 Agreements (relevant for R2D/D2R signals including timing acquisition and synchronization
RAN1#120 (Athens, Greece, Feb 17th – 21st, 2025)
SIP related Agreements
Agreement
For the SIP of R-TAS, for providing the start of the R2D transmission, one single design based on Option 1 is supported and further down-selection to be done among Alt 1 and Alt 2 :
Option 1: ON-OFF transmission with following alternatives:
Alt 1: A single ON-OFF transmission with pre-defined duration for each of the ON-OFF, where ON and OFF may have same or different durations
Continue discussion to clarify the duration of each of the ON and OFF
Continue discussion to list the different candidate proposals under Alt1
Alt 2: A multi-ON-OFF transmission with pre-defined duration for each of the ON(s)-OFF(s), where different ON and different OFF may have same or different durations and different parts may have same or different duration
Continue discussion to clarify the duration of each of the ON and OFF
Continue discussion to list the different candidate proposals under Alt2
Only a single fixed value for entire duration of SIP of R-TAS is supported, which is independent of the value of “M” used in CAP and PRDCH
Note: Specific design and duration for SIP of R-TAS are further discussed, and companies are encouraged to evaluate the designs in terms of target MDR of [10%] for a FAR up to [1%] and at least following assumptions are used:
MDR refers to the probability that SIP is not detected when it was actually transmitted
FAR probability that the receiver incorrectly detects SIP when SIP was not transmitted
Energy/edge detection-based method is the baseline assumption for evaluation purpose
Continue discussion on necessary details for simulation assumptions

Agreement
For the SIP of R-TAS, down-select among the following candidates:
Alt 1 (Single ON-OFF transmission)
Alt 1-1: ON followed by OFF with same duration for both
Alt 1-2: ON followed by OFF with a duration ratio of 1:[2,3]
Alt 1-3: ON followed by OFF with a duration ratio of [2,3]:1
Alt 2 (Multi-ON-OFF transmission)
Alt 2-1: A number of repetition instances of Alt 1-1 or Alt 1-2 or Alt 1-3
Alt 2-2: ON-OFF-ON (duration of ON and OFF can be different)
Alt 2-3: OFF-ON-OFF (duration of ON and OFF can be different)
Alt 2-4: Combination of single instance of Alt 1-1 and single instance of Alt 1-2
For the evaluation purpose, for both options, candidate values related to duration are considered:
Entire duration of SIP: 1/2 OFDM symbol duration or 1 OFDM symbol duration (including clarifying whether OFDM symbol duration includes CP); additional durations can be considered and reported by companies with justification 
Companies to report the exact duration(s) for ON or OFF
Companies are encouraged to report at least the following details for the evaluations:
Baseline assumption is that RF transmission is not present; companies can report other consideration
For FAR calculation, whether noise and/or PRDCH transmission is considered
Details on threshold detection method including whether/how threshold detection training is used based on the proposed design alternative or not
BW assumptions for RF-ED and BB-LPF
Target MDR of up to 1% for FAR of up to [1%, 10%]

CAP related Agreements
Agreement
For the CAP of R-TAS, the starting chip has a different voltage level compared to the end of the SIP of R-TAS.

Agreement
For the design of the CAP of R-TAS, at least 2 transition edges in same direction are included, i.e. at least two transitions from “OFF” chip to “ON” chip or two transitions from “ON” chip to “OFF” chip.

Agreement
For the CAP of R-TAS:
Candidate values for maximum duration of CAP to be further down-selected to one value from : 1.5 OFDM symbol duration, 2 OFDM symbol duration, 3 OFDM symbol duration
For option 1 for CAP of R-TAS from TR 38.769, maximum duration is applicable to minimum value of M to be supported, and the CAP duration becomes shorter with increasing value of M
FFS: whether the number of ON/OFF transmissions in the CAP is fixed or not fixed
For option 2 for CAP of R-TAS from TR 38.769, maximum duration is the only (constant) duration that is applicable for all the M values to be supported
Down-selection between option 1 and option 2 for CAP of R-TAS from TR 38.769 by RAN1#120-bis
FFS: Values of M to be supported


R2D Midamble related Agreement
Agreement
R2D transmission does not include a midamble.

D2R X-amables related Agreement
Agreement
For D2R preamble design, the functionalities of timing acquisition, SFO estimation/time tracking and channel estimation should be supported
For D2R midamble design, the functionalities of SFO estimation/time tracking and channel estimation should be supported
D2R midamble can be transmitted at the end of the PDRCH transmission. If it is at the end, it is not designed for being used for indicating the end of PDRCH transmission
FFS: condition(s) and/or indication where the D2R midamble is present or not

Agreement
For D2R x-ambles:
Following is considered as the types for base sequence and to be further down-selected:
Option 1: M-sequence 
Option 2: Golay sequence
Note: Above doesn’t preclude an additional part for preamble, e.g. with ON and/or OFF transmission, if needed/supported
FFS: Whether/what multiple sequences (using same base sequence type) are supported
Note: This in no way implies that there is going to be CDMA between D2R x-ambles
For evaluation purpose, companies are encouraged to consider following:
Performance at least in terms of autocorrelation/cross-correlation property, SFO estimation/Timing accuracy, SNR for target PDRCH BLER of [1%, 10%]
Report presence and time-domain resource(s) x-ambles
Report sequence type(s) and length(s) for x-ambles
Following format can be considered for reporting the evaluation results

TDoc file conclusion not found

3GPP TSG RAN WG1 #120bis		        					     R1-2502610
Wuhan, China, April 7th – 11th, 2025

Agenda Item:	9.4.3
Source: 	Moderator (Apple)
Title:	FL Summary#2 on timing acquisition & synchronization for Ambient IoT
Document for:	Discussion & Decision

Introduction & Work-Plan for RAN1#120bis
This document provides the feature lead summary on the offline discussions/inputs/proposals for AI 9.4.3 timing acquisition and synchronization signals for R2D and D2R for ambient IoT WI during RAN1#120bis. 


Contact Information
Please consider providing your company name, your name and email address to be able to reach for any potential offline discussions/contact regarding AI 9.4.3 on timing acquisition and synchronization for ambient IoT.





Work Plan
Following is my plan for this meeting for this agenda for RAN1#120bis:

SIP of R-TAS
For SIP, at least down-select to not more than 2 alternatives from the agreed alternatives in RAN1#120 considering technical justification and evaluation analysis. Also agree on fixed single duration for the down-selected alternatives. 
CAP of R-TAS
For CAP, plan is to down-select between option 1 and option 2 (as agreed in last meeting) and then also limit the number of specific CAP pattern design and hopefully have not more than 2 patterns for further down-selection by end of this meeting
R2D Postamble aspects
For indicating the end of PRDCH transmission, at least down-select to whether R2D control information is used, or Manchester coding rule violation is used. 
D2R X-amble
For D2R x-amble, first priority is to agree on the sequence type and length(s). In addition, for midambles, agree on the method/signaling that allows the device to determine the number of midamble(s) and their location within the PDRCH.

SIP of R-TAS

[Closed] 1st Discussion Round 
For the SIP of R-TAS, following aspects have been discussed in the contributions following the agreements and discussion from RAN1#120:
Preferred alternative(s) for SIP design
Duration of SIP
For comparison among different alternatives, companies also provided simulations results/analysis to justify their preferred alternative(s). In addition, a few companies also discussed other aspects including M value/chip duration for the SIP, whether/what impact of CP insertion on SIP. In the following table, companies’ views on their preferred alternatives along with justification and performance analysis, if provided are summarized. 

Table 2-1: Summary of views on alternatives for SIP



FL Observations:
28 companies provided their views on SIP of R-TAS
11 companies provided evaluation in terms of target MDR and FAR with most of the companies considering MDR <= 1% for a target FAR <= 1%
For Alt 1-1, 3 companies provided their preference with ON-OFF ratio of 1:1 & total duration spanning 1 OFDM symbol and only 1 company [China Telecom] showed that it performs better than other alternatives while it satisfies the target MDR and FAR. On the other hand, 4 companies [Futurewei, Xiaomi, Samsung, Huawei] showed other alternatives that perform better than Alt 1-1.
For Alt 1-2, overall, 8 companies provided their preference for at least some variant of Alt 1-2 and 6 companies [Futurewei, ZTE, Spreadtrum, Oppo, Xiaomi, Qualcomm] showed that at least some variant of Alt 1-2 performs better than other alternatives including Alt 1-1, Alt 1-3, Alt 2-2, Alt 2-4. On the other hand, 4 companies [China Telecom, Samsung, CMCC, Huawei] showed other alternatives that performs better than Alt 1-2. Within Alt 1-2:
2 companies prefer ON-OFF with ratio of 1:3 & total duration spanning 0.5 OFDM symbol
5 companies prefer ON-OFF with ratio of 1:3 & total duration spanning 1 OFDM symbol
3 companies prefer ON-OFF with ratio of 1:2 & total duration spanning 1 OFDM symbol
For Alt 1-3, overall, 6 companies provided their preference for at least some variant of Alt 1-3 and only 1 company [Samsung] showed that at least some variant of Alt 1-3 performs better than other alternatives including Alt 1-1, Alt 1-2. On the other hand, 6 companies [China Telecom, ZTE, Futurewei, Oppo, Qualcomm, Huawei] showed other alternatives that performs better than Alt 1-3. Within Alt 1-3:
4 companies prefer ON-OFF with ratio of 2:1 & total duration spanning 1 OFDM symbol
5 companies prefer ON-OFF with ratio of 3:1 & total duration spanning 1 OFDM symbol
For Alt 2-1, 7 companies provided their preference with ON-OFF ratio of 1:1 & total duration spanning 1 OFDM symbol and no company showed that it performs better than other alternatives. On the other hand, 2 companies [China Telecom, Vivo] showed other alternatives that perform better than Alt 2-1.
For Alt 2-2, overall, 6 companies provided their preference for at least some variant of Alt 2-2 and only 1 company [Vivo] showed that at least some variant of Alt 2-2 performs better than other alternatives including Alt 2-1, Alt 2-4. On the other hand, 2 companies [China Telecom, ZTE] showed other alternatives that performs better than Alt 2-2. Within Alt 2-2:
1 company prefer ON-OFF-ON with ratio of 1:1:4 & total duration spanning 1 OFDM symbol
2 companies prefer ON-OFF-ON with ratio of 1:4:1 & total duration spanning 1 OFDM symbol
3 companies prefer total duration spanning 1 OFDM symbol without any specific consideration on ration of ON/OFF
For Alt 2-3, overall, 3 companies provided their preference for at least some variant of Alt 2-3 and no company showed that it performs better than other alternatives including. On the other hand, 2 companies [China Telecom, Huawei] showed other alternatives that performs better than Alt 2-3. Within Alt 2-3:
2 companies prefer OFF-ON-OFF with ratio of 1:1:4 & total duration spanning 1 OFDM symbol
1 company prefer OFF-ON-OFF with ratio of 1:4:1 & total duration spanning 1 OFDM symbol
1 company prefer OFF-ON-OFF with total duration spanning 1 OFDM symbol
For Alt 2-4, 4 companies provided their preference with ON-OFF-ON-OFF ratio of 1:1:1:3 & total duration spanning 0.5 OFDM symbol and 2 companies [CMCC, Huawei] showed that it performs better than other alternatives. On the other hand, 2 companies [ZTE, Vivo] showed other alternatives that perform better than Alt 2-4.
A few companies discussed M value for SIP and whether/how the selection of M may impact the SIP detection considering CP insertion. However, based on discussions during RAN1#120, we considered discussing in terms of OFDM symbol duration and avoid reference M value. Therefore, with that understating, we should not try to reopen the discussion of reference M value but rather have discussion in terms of OFDM symbol duration. Eventually, it will be up to the specification editor on how to capture the SIP duration and corresponding ON(s) and OFF(s). Furthermore, at least 1 company raised the issue that for chip duration corresponding to M < 6, minimum PRB of 2 will be needed. However, another company clarified that we don’t need to associate the mapping of M to minimum number of PRBs for SIP and device should be able to adjust its LPF between SIP reception and follow-up R2D reception. 
Another aspect that a few discussed is whether/what impact CP might have on SIP detection. One company pointed out that chip duration corresponding to higher value of M map impact SIP detection, while 3 companies pointed out that CP impact is negligible on SIP and should not be considered for specific optimization, especially considering that all companies considered SIP duration of not more than 1 OFDM symbol

Based on above observations from companies’ contributions, as a starting point, FL proposal would be to down-select to Alt 1-2 and Alt 2-4. Main rationale behind these two alternatives is not just majority preference but evaluations showing that these two alternatives at least perform better within alternative set 1 and set 2, respectively. Further focused discussion on Alt 1-2 and Alt 2-4 can be further done, if we down-select these 2 options. Therefore, accordingly Proposal 2-1 is provided below:

(Updated) Proposal 2-1 
For the pattern of SIP of R-TAS, only following 2 alternatives are considered for further down-selection to one alternative:
Alt 1-2: ON-OFF with a ratio of 1:3 and with following total SIP duration to be further down-selected [during RAN1#120bis]:
Option 1: 0.5 OFDM symbol duration
Option 2: 1 OFDM symbol duration
Alt 2-4: ON-OFF-ON-OFF with a ratio of 1:1:1:3 and with following total SIP duration to be further down-selected [during RAN1#120bis]:
Option 1: 0.5 OFDM symbol duration
Option 2: 1 OFDM symbol duration





CAP of R-TAS


[Closed] 1st Discussion Round 
For the CAP of R-TAS, following aspects have been discussed in the contributions following the agreements and discussion from RAN1#120:
Preferred option in terms of variable or fixed duration for CAP design
Preferred exact CAP pattern including M values/duration and CP insertion related aspects

Table 3-1: Summary of views between option 1 and option 2 for CAP




Table 3-2: Summary of views on exact CAP pattern


FL observations
Overall, 30 companies provided their views on CAP of R-TAS
26 companies provided their preference relation to option 1 and option 2 from the TR and 17 companies indicated their preference for option 1, while 8 companies indicated their preference for option 2. 1 company indicated the preference to support and combine both options. Based on the agreement from RAN1#120bis, one option needs to be down-selected in this meeting. From the proponents of option 1, 3 companies provided evaluations and showed that option 1 is able to achieve target FDR, as long as there is sufficient gap between different M values. On the other hand, from the proponents of option 2, 2 companies provided evaluations and showed that at least for higher M values, option 2 is needed. 
Among specific CAP patterns, 5 alternatives are provided. Among the 5 alternatives, majority shows that either 3 chip with alternate ON/OFF or 4 chips with alternate ON/OFF can work well for the M values. Alt 3, 4, 5 with more chips have been proposed only by 1-2 companies each and don’t offer any significant gain compared to Alt 1 and Alt 2 and rather increase the overhead, especially for lower M values
Based on above, proposal 3-1 and proposal 3-2 are provided.



(Updated) Proposal 3-1
For CAP of R-TAS, option 1 for CAP of R-TAS from TR 38.769 is adopted with following details:
CAP duration becomes shorter with increasing value of M
FFS: exact relation between duration of CAP and M values



Proposal 3-2 
For the pattern of CAP of R-TAS, only following 2 alternatives are considered for further down-selection to one alternative:
Alt 1: 3 chips with ON-OFF-ON pattern 
Alt 2: 4 chips with ON-OFF-ON-OFF pattern
Based on the minimum value of M that will be agreed in agenda 9.4.2, corresponding maximum duration of CAP will be selected from the candidate set of {1.5 OFDM symbols, 2 OFDM symbols, 3 OFDM symbols}


[Closed] 2nd Discussion Round 
[Closed] Proposal 3-3
For CAP of R-TAS, following is adopted:
ON-OFF-ON-OFFON pattern with Option 1 for CAP of R-TAS from TR 38.769 is adopted where the CAP duration becomes proportionally shorter with increasing value of M, i.e. if for , duration is  OFDM symbol long, then for , duration is  OFDM symbol long


Proposal 3-3b
For CAP of R-TAS, following is adopted:
ON-OFF-ON-OFFON pattern with Option 1 for CAP of R-TAS from TR 38.769 is adopted where the CAP duration becomes proportionally shorter with increasing value of M, i.e. if for , duration is  OFDM symbol long, then for , duration is  OFDM symbol long
Note: Duration without CP insertion is considered above, with CP insertion, the total duration may not be exactly proportional
Only following two alternatives for CAP pattern  are considered for further down-selection to one alternative:
Alt 1: ON-OFF-ON-OFF
Alt 2: ON-OFF-ON





R2D Postamble



[Closed] 1st Discussion Round 
FL observations
24 companies provided their views in terms of different options for indicating the end of PRDCH transmission
Overall, 11 companies prefer indicating the end of PRDCH transmission based on some unique pattern that violate Manchester coding rule:
7 companies prefer to explicitly specify postamble design
4 companies prefer to not specify postamble design and leave it up to the reader implementation to violate the Manchester coding rule
8 companies prefer indication the end of PRDCH transmission based on R2D control information. Among the 8 companies there are still different views on whether R2D control information is L1 or not
3 companies consider supporting combination of the above 2 options
From FL perspective, all the options work and doesn’t necessarily have any significant performance difference. However, for R2D control indication based method, it is still not clear whether L1 R2D control information is used or not. Based on the proponents, the views are still diverging. If L1 R2D control information is used, then it needs further discussion on separate or joint CRC attachment. On the other hand, for Manchester coding rule based option, in principle companies have common understanding that violation, for example could be based on transmission of three high or three low voltages, i.e. ON-ON-ON or OFF-OFF-OFF. But the main divergence is whether it needs to be specified or not.


Based on above, proposal 4-1 is provided to take the majority view of indicating end of PRDCH transmission based on violation of Manchester coding rule and also with the consideration that with this option, no further discussion is needed on details related to R2D control information and CRC. For method based on violation of Manchester coding rule, furthermore, it can be discussed whether/what postamble needs to be specified or not. If companies cannot converge whether/what postamble needs to be specified or not, then implementation based solution can be adopted.


(Updated) Proposal 4-1 
Take either Alt 1 or Al 2:
Alt 1: For indicating the end of PRDCH transmission, Manchester coding rule violation corresponding to M value is adopted
FFS: whether the Manchester coding rule violation is based on specified postamble pattern or based on device detecting such violation based on reader’s implementation
Note: If companies cannot converge on whether/what postamble is specified or not, then device detecting such violation based on reader’s implementation will be adopted
Note: If no postamble is specified, then from timing perspective, end of last chip of PRDCH is the end of R2D transmission
Alt 2: There is no consensus to specify R2D postamble



[Closed] 2nd Discussion Round 
[Closed] Proposal 4-1b 
At the end of PRDCH, Manchester coding rule violation corresponding to M value is applied by reader’s implementation that can be used by the device to determine the end of PRDCH transmission
Note: From timing perspective, end of last chip of PRDCH is the end of R2D transmission



Proposal 4-1c 
At the end of PRDCH, Manchester coding rule violation corresponding to M value is applied by reader’s implementation that can be used by the device to determine the end of PRDCH transmission
Note: From timing perspective, end of last chip of PRDCH is the end of R2D transmission



Proposed Conclusion 4-2 
There is no consensus to specify R2D postamble




D2R X-ambles (including preamble & midamble)


[Closed] 1st Discussion Round 
Base Sequence Types: First aspect discussed by companies is related to the preference between the two sequence types including the length that were agreed in RAN1#120bis to be further down-selected and corresponding details including sequence length and generation. In Table 5-1 below, the companies’ views are summarized and also including performance analysis, if provided

Table 5-1: Summary of views between M-sequence and Golay Sequence types for D2R X-ambles



Multiple Preamble lengths: Second aspect discussed by companies is related to the number of preamble types in terms of different lengths. In Table 5-2 below, the companies’ views are summarized and also including performance analysis, if provided

Table 5-2: Summary of views on D2R preamble lengths including long and short preamble



# of Preamble Sequences (with same base sequence): Third aspect discussed by companies is related to the number of sequences of same lengths for preamble/midamble. In Table 5-3 below, the companies’ views are summarized and also including performance analysis, if provided

Table 5-3: Summary of views on number of preamble sequences



1- or 2- part Preamble: Fourth aspect is related to 1-part or 2-part preamble design. In Table 5-4 below, the companies’ views are summarized and also including performance analysis, if provided

Table 5-4: Summary of views on number of parts of preamble



Midamble Configuration: Fifth aspect is related to presence/location of D2R x-ambles including whether pre-defined rule and/or explicit signaling by reader is supported. In Table 5-5 below, the companies’ views are summarized and also including performance analysis, if provided

Table 5-5: Summary of views on midamble configuration





Midamble Sequence: Fifth aspect is related to whether midamble sequence is same as preamble or not. In Table 5-6 below, the companies’ views are summarized and also including performance analysis, if provided

Table 5-6: Summary of views on midamble sequence



FL observations
In terms of the base sequence type, all the companies that provided simulation results showed that the performance is similar with both Golay sequence and M sequence. And in terms of preference, almost there is equal number of companies supporting both options. For M sequence, it was pointed out by few companies that the sequence generation is perhaps simpler compared to Golay sequence. Therefore, FL proposal is to adopt M sequence for D2R x-ambles. 

In terms of preamble types, at least 11 companies prefer to support two preamble formats including short preamble format and long preamble format. Multiple companies pointed out the benefit of supporting the two formats and also, one company demonstrated the benefit of using two lengths. In terms of sequence length, 32(-1) bits is proposed by majority of company. For the shorter format, it was shown that 8(-1) bits have reasonable performance difference, while 16-(1) bits have almost similar performance. Therefore, FL proposes to support both short and long preamble format, with length 8(-1) bits and 32(-1) bits, respectively. Also, all but 2 companies considered same sequence for both midamble and preamble as both serve similar functionality and this would require less specification effort. Therefore, same sequence for both is also proposed by FL. In terms of 1-part or 2-part format for preamble, 4 companies prefer to have 2-part for coarse and fine synchronization, respectively. On the other hand, 2 companies explicitly proposed not to support 2-part preamble. One company showed in evaluations that as long as preamble and midamble are designed optimally in terms of length and sequence, there is not benefit from 2-part preamble. In order to get views from more companies, FL will ask companies to provide inputs on their preference

Regarding the configuration of midambles, ~13 companies prefer explicit indication by the network, while 8 companies prefer pre-defined rule to determine number of midambles and location. In terms of pre-defined rule, majority consider supporting TBS based midamble determination. Furthermore, there are 7 companies prefer a combination of pre-defined rule and explicit indication by network. From FL perspective, all solutions work, but it seems there is a quite good number of companies that prefer additional control to reader in terms of at least the presence and number of midamble due to potentially dynamic change in channel conditions and/or reader-specific implementation. Therefore, FL proposal would be to consider the combined option where the number of midambles can be explicitly indicated by the reader to the device via R2D control information and the position of midambles can be at least pre-defined. 


Based on above summary of views from companies on D2R x-ambles, following proposals are provided




(Updated) Proposal 5-1 
For D2R, 1-part preamble and 1-part midamble, base sequence is generated from M-sequence, where the length of the sequence is 
Value(s) of n = 5, 3
Long preamble is generated based on n = 5
Short preamble is generated based on following n values to be down-selected to single value
Alt 1: n = 3
Alt 2: n = 4
FFS: signaling/application of the two preambles
Same base sequence is applied for both preamble and midamble
Preamble immediately precedes the PDRCH without any gap


[Closed] Question 5-1 
Do you support 1-part or 2-part preamble, where for 2-part preamble, 1st part is based on 0s on 1s without Manchester coding and 2nd part is based on the base sequence (as proposed in proposal 5-1)?


(Updated) Proposal 5-2
For D2R midamble, following two alternatives are considered for further down-selection in RAN1#120bis for the determining the presence and location of midamble(s) at the device:
Alt 1: Reader explicitly indicates the number of midambles to be applied by the device for PDRCH transmission via R2D control information
FFS: details of signaling
Based on the indicated number of midambles, a pre-defined rule is applied at the device to determine the exact location of each of the midamble(s) 
FFS: details of pre-defined rule for determining the exact location of midamble(s) 
Alt 2: Reader explicitly indicates the interval between consecutive x-ambles
FFS: details of signaling
Note: These alternatives don’t preclude support of no midamble




[Closed] 2nd Discussion Round 
[Closed] Proposal 5-1b 
For D2R, 1-part preamble and 1-part midamble, base sequence is generated from M-sequence, where the length of the sequence is 
Value(s) of n
Long preamble is generated based on n = 5
Short preamble is generated based on following n values to be down-selected to single value
Alt 1: n = 3
Alt 2: n = 4
FFS: signaling/application of the two preambles
Same base sequence is applied for both preamble and midamble
Preamble immediately precedes the PDRCH without any gap



Proposal 5-1c 
For D2R preamble/midamble, base sequence is generated from m-sequence, where the length of the sequence is 
Value(s) of n
Long preamble/midamble is generated based on n = 5
Short preamble/midamble is generated based on following n values to be down-selected to single value
Alt 1: n = 3
Alt 2: n = 4
Only 1-part preamble/midamble are supported for D2R
Preamble immediately precedes the PDRCH without any gap
If both long and short preamble/midamble are supported, then reader separately indicates 
For preamble, whether long preamble or short preamble is applied
For midamble, whether long midamble or short midamble is applied
For the same length of preamble and midamble, same sequence is supported


(Updated) Proposal 5-2 
For indicating the interval between consecutive midambles, and between the preamble and the first midamble, via R2D control information, following is adopted:
Unit of interval
Alt 1: Number of “pairs of chips”
Alt 2: Number of bits
FFS: the candidate values in terms of the unit of interval



Proposals for offline sessions
1st offline session (Monday, April 7, 2025)
(Updated) Proposal 5-2
For D2R midamble, following two alternatives are considered for further down-selection in RAN1#120bis for the determining the presence and location of midamble(s) at the device:
Alt 1: Reader explicitly indicates the number of midambles to be applied by the device for PDRCH transmission via R2D control information
FFS: details of exact signaling to be discussed under agenda 9.4.4
Based on the indicated number of midambles, pre-defined rule(s) are applied at the device to determine the exact location of each of the midamble(s) 
FFS: details of pre-defined rule for determining the exact location of midamble(s)
Alt 2: Reader explicitly indicates the interval between consecutive x-ambles
FFS: details of exact signaling
Note: These alternatives don’t preclude support of no midamble

(Updated) Proposal 3-1
For CAP of R-TAS, option 1 for CAP of R-TAS from TR 38.769 is adopted with following details:
CAP duration becomes shorter with increasing value of M
FFS: Exact relation between duration of CAP and M values

(Updated) Proposal 2-1 
For the pattern of SIP of R-TAS, only following 2 alternatives are considered for further down-selection to one alternative:
Alt 1-2: ON-OFF with a ratio of 1:3 and with following total SIP duration to be further down-selected during RAN1#120bis:
Option 1: 0.5 OFDM symbol duration
Option 2: 1 OFDM symbol duration
Alt 2-4: ON-OFF-ON-OFF with a ratio of 1:1:1:3 and with following total SIP duration to be further down-selected during RAN1#120bis:
Option 1: 0.5 OFDM symbol duration
Option 2: 1 OFDM symbol duration

Proposal 4-1 
For indicating the end of PRDCH transmission, Manchester coding rule violation corresponding to M value is adopted
FFS: whether the Manchester coding rule violation is based on specified postamble pattern or based on device detecting such violation based on reader’s implementation
Note: If companies cannot converge on whether/what postamble is specified or not, then device detecting such violation based on reader’s implementation will be adopted
Note: If no postamble is specified, then from timing perspective, end of last chip of PRDCH is the end of R2D transmission


(Updated) Proposal 5-1 
For D2R, 1-part preamble and 1-part midamble, base sequence is generated from M-sequence, where the length of the sequence is 
Value(s) of n = 5, 3
Long preamble is generated based on n = 5
Short preamble is generated based on following n values to be down-selected to single value
Alt 1: n = 3
Alt 2: n = 4
FFS: signaling/application of the two preambles
Same base sequence is applied for both preamble and midamble
Preamble immediately precedes the PDRCH without any gap

Proposal 3-2 
For the pattern of CAP of R-TAS, only following 2 alternatives are considered for further down-selection to one alternative:
Alt 1: 3 chips with ON-OFF-ON pattern 
Alt 2: 4 chips with ON-OFF-ON-OFF pattern
Based on the minimum value of M that will be agreed in agenda 9.4.1, corresponding maximum duration of CAP will be selected from the candidate set of {1.5 OFDM symbols, 2 OFDM symbols, 3 OFDM symbols}


2nd offline session (Tuesday, April 8, 2025)
Proposal 5-1b 
For D2R, 1-part preamble and 1-part midamble, base sequence is generated from m-sequence, where the length of the sequence is 
Value(s) of n
Long preamble/midamble is generated based on n = 5
Short preamble/midamble is generated based on following n values to be down-selected to single value
Alt 1: n = 3
Alt 2: n = 4
If both long and short preamble/midamble are supported, reader explicitly indicates it
Same base sequence is applied for both preamble and midamble
Preamble immediately precedes the PDRCH without any gap



Proposal 3-3a
For CAP of R-TAS, following is adopted:
ON-OFF-ON-OFFON pattern with Option 1 for CAP of R-TAS from TR 38.769 is adopted where the CAP duration becomes proportionally shorter with increasing value of M, i.e. if for , duration is  OFDM symbol long, then for , duration is  OFDM symbol long
Note: Duration without CP insertion is considered above, with CP insertion the total duration may not be exactly proportional

Proposal 4-1c 
At the end of PRDCH, Manchester coding rule violation corresponding to M value is applied by reader’s implementation that can be used by the device to determine the end of PRDCH transmission
Note: From timing perspective, end of last chip of PRDCH is the end of R2D transmission


Proposed Conclusion 4-2 
There is no consensus to specify R2D postamble


3rd offline session (Wednesday, April 9, 2025)
Proposal 5-1c 
For D2R preamble/midamble, base sequence is generated from m-sequence, where the length of the sequence is 
Value(s) of n
Long preamble/midamble is generated based on n = 5
Short preamble/midamble is generated based on following n values to be down-selected to single value
Alt 1: n = 3
Alt 2: n = 4
Only 1-part preamble/midamble are supported for D2R
Preamble immediately precedes the PDRCH without any gap
Reader explicitly indicates the following cases to the device
Case of indicating short preamble and short midamble is supported
Case of indicating long preamble and long midamble is supported
[Case of indicating long preamble and short midamble is supported]
[For the same length of preamble and midamble, same sequence is supported]


Proposal 3-3b
For CAP of R-TAS, following is adopted:
ON-OFF-ON-OFFON pattern with Option 1 for CAP of R-TAS from TR 38.769 is adopted where the CAP duration becomes proportionally shorter with increasing value of M, i.e. if for , duration is  OFDM symbol long, then for , duration is  OFDM symbol long
Note: Duration without CP insertion is considered above, with CP insertion, the total duration may not be exactly proportional
Only following two alternatives for CAP pattern  are considered for further down-selection to one alternative:
Alt 1: ON-OFF-ON-OFF
Alt 2: ON-OFF-ON

(Updated) Proposal 5-2 
For indicating the interval between consecutive midambles, and between the preamble and the first midamble, via R2D control information, following is adopted:
Unit of interval
Alt 2: Number of bits after FEC (if FEC is applied) and repetition
FFS: the candidate values in terms of the unit of interval

Proposal 4-1c 
At the end of PRDCH, Manchester coding rule violation corresponding to M value is applied by reader’s implementation that can be used by the device to determine the end of PRDCH transmission
Note: From timing perspective, end of last chip of PRDCH is the end of R2D transmission

Proposed Conclusion 4-2 
There is no consensus to specify R2D postamble


Proposals for online session
1st online session (Tuesday, April 8, 2025)
Proposal 5-2a
For D2R midamble, following two alternatives are considered for further down-selection in RAN1#120bis for determining the presence and location of midamble(s) at the device:
Alt 1: Reader explicitly indicates the number of midambles to be applied by the device for PDRCH transmission via R2D control information
FFS: details of signaling
Based on the indicated number of midambles, pre-defined rule(s) are applied at the device to determine the exact location of each of the midamble(s) 
FFS: details of pre-defined rule for determining the exact location of midamble(s) 
Alt 2: Reader explicitly indicates the interval between consecutive x-ambles
FFS: details of signaling and any pre-define rule(s), if needed
Note: These alternatives don’t preclude the support of no midamble

Proposal 2-1a
For the pattern of SIP of R-TAS, only following 2 alternatives are considered for further down-selection to one alternative:
Alt 1-2: ON-OFF with a ratio of 1:3 and with following total SIP duration to be further down-selected [during RAN1#120bis]:
Option 1: 0.5 OFDM symbol duration
Note: In this option, the chip duration corresponds to M = 8, but this doesn’t imply support of such reference M value for SIP
Option 2: 1 OFDM symbol duration
Note: In this option, the chip duration corresponds to M = 4, but this doesn’t imply support of such reference M value for SIP
Alt 2-4: ON-OFF-ON-OFF with a ratio of 1:1:1:3 and with following total SIP duration to be further down-selected [during RAN1#120bis]:
Option 1: 0.5 OFDM symbol duration
Note: In this option, the chip duration corresponds to M = 12, but this doesn’t imply support of such reference M value for SIP
Option 2: 1 OFDM symbol duration
Note: In this option, the chip duration corresponds to M = 6, but this doesn’t imply support of such reference M value for SIP

Proposal 3-1a
For CAP of R-TAS, option 1 for CAP of R-TAS from TR 38.769 is adopted with following details:
CAP duration becomes shorter with increasing value of M
FFS: exact relation between duration of CAP and M values including {2,4,6,24}

Proposal 3-2a 
For the pattern of CAP of R-TAS, only following 2 alternatives are considered [for further down-selection to one alternative]:
Alt 1: 3 chips with ON-OFF-ON pattern 
For lowest M  = 2, maximum CAP duration will be 1.5 OFDM symbols with this alternative
Alt 2: 4 chips with ON-OFF-ON-OFF pattern
For lowest M  = 2, maximum CAP duration will be 2 OFDM symbols with this alternative
Based on the minimum value of M that will be agreed in agenda 9.4.2, corresponding maximum duration of CAP will be selected from the candidate set of {1.5 OFDM symbols, 2 OFDM symbols, 3 OFDM symbols}


Proposal 5-1a 
For D2R, 1-part preamble and 1-part midamble, base sequence is generated from M-sequence, where the length of the sequence is 
Value(s) of n = 5, 3
Long preamble is generated based on n = 5
Short preamble is generated based on following n values to be down-selected to single value
Alt 1: n = 3
Alt 2: n = 4
FFS: signaling/application of the two preambles
Same base sequence is applied for both preamble and midamble
Preamble immediately precedes the PDRCH without any gap



Proposal 4-1a 
Take either Alt 1 or Al 2:

Alt 1: For indicating the end of PRDCH transmission, Manchester coding rule violation corresponding to M value is adopted
FFS: whether the Manchester coding rule violation is based on specified postamble pattern or based on device detecting such violation based on reader’s implementation
Note: If companies cannot converge on whether/what postamble is specified or not, then device detecting such violation based on reader’s implementation will be adopted
Note: If no postamble is specified, then from timing perspective, end of last chip of PRDCH is the end of R2D transmission
Alt 2: There is no consensus to specify R2D postamble





















2nd online session (Wednesday, April 9, 2025)

Proposal 5-1d 
For D2R preamble/midamble, base sequence is generated from m-sequence, where the length of the sequence is 
Value(s) of n
Long preamble/midamble is generated based on n = 5
Short preamble/midamble is generated based on following n values to be down-selected to single value
Alt 1: n = 3
Alt 2: n = 4
Only 1-part preamble/midamble are supported for D2R
Preamble immediately precedes the PDRCH without any gap
If both long and short preamble and midamble (if present) are supported, then following cases are supported and reader explicitly  indicates one of the following cases for PDRCH:
Short preamble and short midamble 
Long preamble and long midamble 
[Long preamble and short midamble]
[For the same length of preamble and midamble, same sequence is supported]



Proposal 3-3c
For CAP of R-TAS, following is adopted:
Option 1 for CAP of R-TAS from TR 38.769 is adopted where the CAP duration becomes proportionally shorter with increasing value of M, i.e. if for , duration is  OFDM symbol long, then for , duration is  OFDM symbol long
Note: Duration without CP insertion is considered above, with CP insertion, the total duration may not be exactly proportional
Only following two alternatives for CAP pattern  are considered for further down-selection to one alternative:
Alt 1: ON-OFF-ON-OFF
Alt 2: ON-OFF-ON


Proposal 5-2a 
For indicating the interval between consecutive midambles, and between the preamble and the first midamble, via R2D control information, following is adopted:
Unit of interval is number of bits after FEC (if FEC is applied) and repetition
FFS: the candidate values in terms of the unit of interval



Proposal 4-1c 
At the end of PRDCH, Manchester coding rule violation corresponding to M value is applied by reader’s implementation that can be used by the device to determine the end of PRDCH transmission
Note: From timing perspective, end of last chip of PRDCH is the end of R2D transmission

Proposed Conclusion 4-2 
There is no consensus to specify R2D postamble


Contributions in RAN1#120bis


Appendix
Revised WID (RP-243326): RAN1 Scope & Objectives 
General Scope
The definitions provided in TR 38.848, TR 38.769, and decisions, etc. made during the Rel-19 SI in RAN WGs are taken into this WI, and the following is the exclusive general scope:
The overall objective shall be to standardize the following Ambient IoT device:
Device 1: ~1 µW peak power consumption, has energy storage, RF envelope detector receiver, initial sampling frequency offset (SFO) up to 10X ppm, neither R2D nor D2R amplification in the device. The device’s D2R transmission is backscattered on a carrier wave provided externally.
Deployment scenario 1 with Topology 1, according to D1T1-B. 
FR1 licensed spectrum in FDD, with R2D in DL spectrum and D2R and CW in UL spectrum.
Spectrum deployment in-band to NR and standalone, with A-IoT BS located indoor.
Traffic types DO-DTT, DT, for rUC1 (indoor inventory) and rUC4 (indoor command). 
Carrier wave transmission for waveform 1 only, without hopping, per the following cases in TR 38.769:
Case 1-4 for D1T1-B
Proximity determination via Solution 1 in TR 38.769 only.
Device (un)availability via Direction 1 in TR 38.769 only.

WGs begin their discussions from the decisions already made in TR 38.769, with the following refinements for the scope: 

The following objectives are set, within the General Scope:
RAN1 scope:
PRDCH and PDRCH, which are the only physical channels in R2D and D2R, respectively.
R2D and D2R signal(s)
Multiplexing/multiple access in R2D is by only TDMA, and in D2R is by only TDMA and FDMA.
R2D supports only OOK-4 modulation, one solution for CP handling. D2R backscattering supports only OOK and BPSK modulations.
R2D transmission supports only the Manchester line code in TR 38.769
D2R transmission supports:
Either the Manchester line code in TR 38.769 or no line code (one to be down-selected); and
A corresponding small frequency shift method according to the options in TR 38.769.
R2D does not support FEC. D2R supports only convolutional code with generator polynomials as per TS 36.212. Applying or not applying the FEC to D2R is specified by ensuring it is under the reader control and applies to all devices targeted by the reader.
PRDCH and PDRCH both support transmission without CRC, and with CRC as per the generator polynomials in TS 38.212 for 6-bit CRC and 16-bit CRC. Cases to use which length of CRC, or no CRC, to be decided in RAN1.
D2R supports physical layer repetition transmission. R2D does not support physical-layer repetition transmission. 
RAN2 scope:
Specify the necessary functions and procedures for an Ambient IoT compact protocol stack and lightweight signalling procedure to enable DO-DTT and DT data transmission:
A-IoT Paging, including subsequent paging for the same service. Support the options that a paging message contains one identifier, and that a paging message contains no identifier. 
Note: RAN2 aims to design a paging message format such that multiple identifiers can be contained in one paging message, for forward compatibility purposes.
A-IoT Random access, including re-access for failure handling. Contention-based and contention-free cases are supported. For the contention-based random access, only Solution 1 (3-step only) is included.
A-IoT data transmission, including data (re-)transmission for failure handling. Segmentation is supported at least in D2R.
Only MAC layer is included
RAN3 scope: 
Specify necessary architectural aspects, and signaling and procedures between A-IoT RAN and A-IoT CN to support the A-IoT functions, assuming an architecture of aggregated gNB, including:
Inventory and command operations
Device location reporting at reader ID granularity
Note: The above A-IoT functions are supported over the existing NG interface, based on architecture(s) defined by RAN3/SA2.
RAN4 scope:
Specify RF requirements for Ambient-IoT BS, device 1, and CW
RF requirements for Type 1-C Ambient-IoT BS
RF requirements for device 1
RF requirements for CW
Specify RRM core requirements for device 1, if necessary
Study and develop OTA test methodology for A-IoT device 1
Consider test methods specified in TR 38.870 as starting point. Take test system reuse, test system complexity and test time into account, when developing test methods suitable for Ambient IoT.
Develop the preliminary Measurement Uncertainty (MU) assessment for the test system
Use band n8 as an example band

Note 1: Coordination with SA2 and SA3 is expected. Updates to the WID objectives should be considered if needed.

Note 2: This WI shall target for an IoT segment well below the existing 3GPP IoT technologies, e.g. NB-IoT, eMTC, RedCap, etc. The WI shall not aim to replace existing 3GPP LPWA technologies.

SI Phase: RAN1 Agreements (relevant for R2D/D2R signals including timing acquisition and synchronization
RAN1#116 (Athens, Greece, February 26th – March 1st, 2024)
Agreement
At least the following time domain frame structure is studied for A-IoT R2D and D2R transmission.
For R2D transmission,
A R2D timing acquisition signal (e.g. R2D preamble) is included at least for timing acquisition and for indicating the start of the R2D transmission in time domain.
For D2R transmission,
A D2R timing acquisition signal (e.g. D2R preamble) is included at least for timing acquisition and for indicating the start of the D2R transmission in time domain.
FFS other necessary component(s), e.g. midamble, postamble, periodic sync signal, control fields, guard period


RAN1#116bis (Changsha, Hunan Province, China, April 15th – April 19th, 2024)

Agreement
To determine or derive the end of PRDCH transmission, study at least following options:  
Option 1: R2D postamble immediately follows the PRDCH to indicate the end of the PRDCH.       
Option 2: Based on R2D control information.

Agreement
For the reader to acquire the end of PDRCH transmission, study at least following options:  
Option 1: D2R postamble immediately follows the PDRCH
Option 2: Based on control information

Agreement
For D2R transmission, study the necessity of midamble at least for the purpose of performing timing/frequency tracking or channel estimation or interference estimation, considering at least the following: 
Modulation and Coding schemes, e.g., data modulation, line/channel coding 
Receiving methods, e.g., coherent or non-coherent
D2R transmission length/packet size
Midamble overhead
Timing/frequency accuracy
Phase accuracy

Agreement
RAN1 study the R2D transmission without midamble as the baseline if Manchester encoding is used.
FFS the necessity for the R2D transmission with midamble if PIE is used. 

Agreement
For the R2D timing acquisition signal immediately preceding the transmission of a physical channel, study a preamble with at least two parts which includes a start-indicator part and a clock-acquisition part, where the start-indicator part immediately precedes the clock-acquisition part:
Start-indicator part provides the start of the R2D transmission
FFS: Details of start-indicator part
Clock-acquisition part provides at least the chip synchronization of the subsequent physical channel transmission
FFS: Details of clock-acquisition part, e.g. structure, encoding, length, etc. 
FFS: Methods to determine chip duration of the subsequent physical channel transmission 
FFS: Other functionalities
Note: the preamble is considered not to be part of a physical channel
FFS: other part(s) of the preamble, if any 
FFS: whether the above clock acquisition is sufficient for all devices
FFS: how to make the preamble compact

Agreement
For D2R, a preamble preceding each PDRCH transmission is studied as the baseline at least for the D2R timing acquisition signal:
Preamble is not part of PDRCH
FFS: Other functionalities of the preamble

Agreement
Reference signals including at least DMRS, PTRS, CSI-RS/TRS, are not further studied for R2D.

Agreement
Reference signals including DMRS, PTRS, SRS, are not further studied for D2R
Note: This doesn’t preclude the possibility to study preamble, midamble, postamble for different purposes, e.g. channel/interference estimation and/or proximity determination


RAN1#117 (Fukuoka City, Fukuoka, Japan, May 20th – 24th, 2024)

Agreement
For the start-indicator part of the R2D time acquisition signal, study the two options below:
Option 1: ON/OFF pattern i.e. high/low voltage transmission 
Option 2: OFF pattern, i.e. low voltage transmission 

Agreement
For R2D, the clock-acquisition part of the R2D time acquisition signal is used to determine the OOK chip duration
FFS: Pattern design to support determination of chip duration


RAN1#118 (Maastricht, NL, August 19th – 23rd,  2024)

Agreement
For each D2R transmission, no separate part for start-indicator is considered for the preamble preceding the PDRCH.

Agreement
For D2R transmission, preamble preceding the PDRCH is studied also for the potential additional functionalities:
SFO estimation
CFO estimation
Channel estimation
Interference estimation
Note: this does not preclude studying the above functionalities by using a midamble and/or postamble, if supported
FFS: Other functionalities, if any

Agreement
For the start-indicator part of the R2D time acquisition signal, ON/OFF pattern i.e. high/low voltage transmission is applied
FFS: length/pattern of ON/OFF.
FFS: when TD2R_min is applicable, whether/how the start-indicator part is included in TD2R_min or not. To be discussed in 9.4.2.2


RAN1#118bis (Hefei, China, October 14th – 18th,  2024)

Agreement
The start indicator part of the R2D time acquisition signal is not included in TD2R_min.

Agreement
The TR will capture the following options, and companies are encouraged to analyze the tradeoffs among the following D2R amble(s) options:
Option 1: D2R preamble only
Option 2: D2R preamble + X midamble(s), where X 1
Option 3: D2R preamble + postamble
Option 4: D2R preamble + Y midamble(s) + postamble, where Y1
For the above options, companies are encouraged to report at least the following:
Purpose(s) of the preamble, midamble and postamble 
Whether companies assume multiple options can be supported


Agreement
For analysing the trade-offs among the D2R amble(s) options, companies can refer to the Table 3.2.4 in section 3.2.4 of R1-2408993 for information.

Agreement
For the clock-acquisition part of the R2D time acquisition signal, following is captured in the TR 38.769:
Clock-acquisition part is based on OOK without line coding and includes rising/falling edges, including at least two rising or at least two falling edges for the device to determine the OOK chip duration

Agreement
For the start-indicator part of the R2D time acquisition signal, for providing the start of the R2D transmission, following is captured in the TR 38.769:
Following options have been studied for the start-indicator part of the R2D time acquisition signal:
Option 1: ON-OFF transmission is considered based on energy/edge detection, and multiple alternatives have been studied including 
Alt 1: A single ON-OFF transmission, i.e. one high-voltage transmission followed by one low-voltage transmission, where ON and OFF may have same or different durations
Alt 2: A multi-ON-OFF transmission, where different ON and different OFF may have same or different durations and different parts may have same or different duration
Option 2: ON-OFF sequence-based design is considered which consists of a pre-defined sequence for detection of start-indicator part based on digital correlation
For both the options, it is observed that a fixed duration for the start-indicator part can be considered, regardless of the value of M used for PRDCH transmissions. 
Miss-detection ratio (MDR), false-alarm ratio (FAR) and detection complexity have been considered for the design of the R2D start indicator part by following companies
It is observed by 1 source [Huawei] that for an FAR of ~0%, the MDR of less than 1% can be achieved with Alt 2 of option 1 (considering 2 ON-OFF transmissions with different durations) and it is also observed that low-complexity and reduced power consumption can be achieved
1 source [ZTE] evaluated Alt 1 of option 1 (considering same duration for ON and OFF) and Alt 2 of option 1 (considering multiple ON-OFF transmissions with same duration) and observed that for an FAR of ~0%, the MDR of less than 1% can be achieved and Alt 1 of option 1 performs better than Alt 2 of option 1. 
1 source [CATT] observed with ON-OFF pattern, that for an FAR of ~0%, the MDR of less than 1% can be achieved with a duration of at least 1 OFDM symbol
1 source [Qualcomm] compares the performance between option 1 and option 2. It shows almost similar coverage range (SNR requirement) for target MDR of 1%. For MDR of 10%, it shows that sequence-based design provides better performance, and it is observed that during the available time, it is feasible for all devices to detect the start-indicator sequence within the power budget. It is further observed that the FAR with sequence-based design can be improved in case of interference scenarios when compared with pattern-based design. 
For both the options, it may be beneficial that the start-indicator part is distinguishable at least from other parts of the R2D transmissions

Agreement
For the clock-acquisition part of the R2D time acquisition signal for OOK chip duration determination, following options are studied:
Option 1: Duration of the clock-acquisition part is variable for different M values, i.e. the duration becomes shorter with increasing value of M
Option 2: Duration of the clock-acquisition part is constant for different M values based on repetition, i.e. repetition factor is increased with increasing value of M to keep the duration constant
FFS: Whether/what restriction on M values for the clock-acquisition part
Note: Other functionalities of clock-acquisition part is a separate discussion

Agreement
For the D2R preamble, binary signal is considered.



RAN1#119 (Orlando, US, Nov 18th – 22nd, 2024)
Agreement
Capture following observations in the TR 38.769, where CFO is assumed to be zero or negligible.
For coherent detection of PDRCH with a payload of 16 bits or 20 bits with 6-bit or 16-bit CRC, using 1/2 Manchester coding and 1/3 or 1/2 convolutional code:
Source [6, ZTE] observed that with up to 10% SFO, ~1kbps data rate, when the same amble(s) overhead is maintained, Option 3 provides comparable performance results to Option 1.
Source [7, Samsung] observed that with up to 10% SFO, ~5kbps data rate, for device 1 and with up to 1% SFO for device 2, the decoding performance with/without midamble are similar
Source [9, vivo] observed that Option 1 is sufficient to achieve 10% and 1% BLER, with no more than 8 SFO hypotheses tested at the reader side.
With up to 10% SFO, ~ 5kbps data rate, the SNR needed to achieve 10% and 1% BLER is similar (~ -2dB and 4 dB) for Option 1, Option 2 of D2R preamble+1midamble and Option 3.
With up to 1% SFO, ~ 5kbps data rate, the SNR needed to achieve 10% and 1% BLER is similar (~ -2.8dB and 3.3dB) for Option 1, Option 2 of D2R preamble+1 midamble and Option 3.
Source [20, OPPO] observed that with perfect SFO estimation, 1kbps data rate and OOK modulation, there is no noticeable performance gain from using midamble(s) and/or postamble for PDRCH transmission, regardless of the payload size.
Source [30, Qualcomm] observed that the required SFO estimation accuracy to achieve 1% and 10% BLER is <5*10^3 ppm
With up to 10% SFO, achieving the required accuracy necessitates more than 20 SFO hypotheses at the reader side for Option 1 and 10 SFO hypotheses are sufficient for Option 3 of D2R preamble + postamble. But for Option 3 reader has to store the received samples and wait for the postamble that is after the end of PDRCH for any of SFO/channel/interference estimation, demodulation, decoding, etc. 
With up to 1% SFO, 4 SFO hypotheses are sufficient for Option 1 to achieve the required accuracy.

For coherent detection of PDRCH with a payload of 96bits with 16-bit CRC (or 6-bit CRC [14, Xiaomi]), using 1/2 Manchester coding and 1/3 or 1/2 convolutional code,
Sources [3, Huawei], [5, CMCC] and [14, xiaomi] observed that Option 1 cannot achieve 10% BLER.
Sources [6, ZTE], [7, Samsung], [9, vivo], [20, OPPO] and [30, QC] observed that Option 1 can achieve 10% BLER.
Sources [3, Huawei], [5, CMCC], [6, ZTE], [7, Samsung], [9, vivo], [14, xiaomi], [16, China Telecom] observed that adding additional amble improves the performance. 
Source [3, Huawei] observed that with up to 10% SFO, 
Option 2 of D2R preamble+ 1 midamble achieves 10% BLER at SNR around -3dB, but cannot achieve 1% BLER.
Option 3 of D2R preamble+ postamble achieves 10% BLER at SNR around -4dB, and can achieve 1% BLER at SNR around 4dB.
Source [5, CMCC] observed that with up to 10% SFO, Option 3 allows reader to precisely search and detect the SFO with 0.03% residual SFO at -3dB SNR TDL-A channel, achieving 10% BLER -2.44dB SNR for ~1 kbps data rate and -2.17 dB for ~6 kbps data rate. Source [5, CMCC] further observed that when the reader adopts same number of SFO hypothesis based on preamble, with 1% SFO, Option 3 can achieve 10% BLER at -4.27 dB SNR for ~1 kbps and at -4.29 dB SNR for ~6 kbps, which provides 1~2 dB performance gain when compared to 10% SFO. 
Source [14, xiaomi] observed that with up to 10% SFO, Option 2 of D2R preamble + 1 midamble, option 3, and option 4 of D2R preamble + 1 midamble+postamble achieve basically the same performance, the SNR for 10% BLER is 5dB for 1.25 kbps data rate.  
Source [6, ZTE] observed that with up to 10% SFO, ~1kbps data rate, and the same amble(s) overhead, Option 3 can provide 1~2 dB, 5dB performance gain for 10% BLER, 1% BLER, respectively, compared to Option 1. Additionally, Option 3 can provide ~1dB, 2dB performance gain for 10% BLER, 1% BLER, respectively, compared to Option 2 of D2R preamble +1 midamble.
Source [7, Samsung] observed that for ~5kbps data rate, compared to option 1, 
For device 1 with up to 10% SFO, Option 2 of D2R preamble + 1 midamble provides ~0.5 dB SNR gain at 10% BLER with TDL-A channel and ~0.9 dB SNR gain with TDL-D channel.
For device 2 with up to 1% SFO, Option 2 of D2R pramble + 1midamble provides ~1 dB SNR gain at 10% BLER with TDL-A channel and ~1.4 dB SNR gain with TDL-D channel.
Source [9, vivo] observed that, 
With up to 10% SFO, ~5.5kbps data rate, Option 1 achieves 10% and 1% BLER at SNR ~0.7dB and 10dB, respectively; Additionally, maintaining the same amble overhead, Option 2 (D2R preamble + 1 midamble) and Option 3 demonstrate similar performance, achieving 10% and 1% BLER at SNR around -1.7dB and 5.2dB, respectively.
With up to 1% SFO, ~5.5kbps data rate, Option 1 achieves 10% and 1% BLER at SNR ~ -1.3dB and 11dB, respectively. Additionally, with the same amble overhead, the SNR difference between Option 2 (D2R preamble+1midamble) and Option 3 for 10% and 1% BLER is less than 1dB, with SNRs ~ -3.1dB to -2.5dB for 10% BLER and ~3.6dB to 4.5dB for 1% BLER.
Source [16, China Telecom] observed that with up to 10% SFO, ~7.5kbps data rate, there is ~6~7dB performance gap at 10% BLER and ~10.5~11.5dB performance gap at 1% BLER between option 2 of D2R preamble+111 midambles and option 1. Note that Source [16, China Telecom] does not use any convolutional code.
Source [20, OPPO] observed that with perfect SFO estimation, 1kbps data rate and OOK modulation, there is no noticeable performance gain from using midamble(s) and/or postamble for PDRCH transmission, regardless of the payload size.
Source [30, Qualcomm] observed that the required SFO estimation accuracy to achieve 1% and 10% BLER is <10^3 ppm. To achieve the required accuracy,
For Option 1, more than 50 SFO hypotheses at reader side are necessary for device with up to 10% SFO and 6 SFO hypotheses are sufficient at reader side for device with up to 1% SFO. 
For Option 3, 10 SFO hypotheses are sufficient for device with up to 10% SFO, but reader has to store the received samples and wait for the postamble that is after the end of PDRCH for any of SFO/channel/interference estimation, demodulation, decoding, etc.

For coherent detection of PDRCH with a payload of 400bits with 16-bit CRC, using 1/2 Manchester coding and 1/3 or 1/2 convolutional code, 
For option 1 of D2R preamble only, 
Sources [3, Huawei], [5, CMCC], [6, ZTE], [8, Spreadtrum], [9, vivo], [14, xiaomi] observed that with up to 10% SFO, 10% BLER cannot be achieved. 
Source [20, OPPO] observed that with perfect SFO estimation, 1kbps data rate and OOK modulation, there is no noticeable performance gain from using midamble(s) and/or postamble for PDRCH transmission.
For other amble options, 
Source [3, Huawei] observed that
With accurate SFO estimation, Option 2 of D2R preamble + 4 midambles can achieve 10% BLER at SNR ~ 2.7dB but cannot achieve 1% BLER.
With up to 10% SFO, Option 3 cannot achieve 10% BLER.
With up to 10% SFO, Option 4 of D2R preamble+2 midambles+postamble achieves 10% BLER at SNR of ~0.25dB; But it cannot achieve 1% BLER. Option 4 of D2R preamble+3 or 4 midambles+postamble, achieves a 10% BLER at an SNR of around -0.2 dB, and achieves 1% BLER at SNR around 9dB or 8dB, respectively.
Source [5, CMCC] observed that with up to 10% SFO, Option 4 of D2R preamble combined with 1 to 4 midambles + postamble, achieves 10% BLER at SNR of 2.5 dB, 1 dB, 0.8 dB, or 0.5 dB, respectively, for a data rate of around 1 kbps.
Source [6, ZTE] observed that with up to 10% SFO, 
Option 3 can provide ~5.5 dB performance gain compared to option 2 of D2R preamble+1midamble for 10% BLER, with the same amble(s) overhead for ~1kbps data rate.
Option 2 of D2R preamble+1midamble cannot achieve 1% BLER for ~1kbps data rate.
Option 4 of the D2R preamble+1 or 2 midamble(s)+postamble, has similar performance, it can achieves a 10% BLER at SNR of -1dB and achieves a 1% BLER at SNR of 6dB and 5dB respectively for ~1kbps data rate.  
Source [8, Spreadtrum] observed that with up to 10% SFO, 
Option 3 of D2R preamble+ postamble cannot achieve 10% BLER for ~7kpbs.  
Option 4 of D2R preamble + 1 midamble + postamble can achieve 10% BLER and 1% BLER at SNR around -6dB and 0 dB, respectively for ~7kpbs data rate.
Source [9, vivo] observed that 
With up to 10% SFO, maintaining the same amble overhead, both Option 2 of D2R preamble+5 midambles and Option 4 of D2R preamble+4 midambles+1 postamble demonstrate similar performance, achieving 10% BLER at SNR ~0.2dB. While for 1% BLER, the SNR for Option 2 and Option 4 is ~ 9.2dB and 12.8dB, respectively for ~5.5kpbs data rate.   
With up to 1% SFO, maintaining the same amble overhead, both Option 2 of D2R preamble+5 midambles and Option 4 of D2R preamble+4 midambles+1postamble demonstrate similar performance, achieving 10% BLER at SNR around -1.2dB. While for 1% BLER, the SNR for Option 2 and Option 4 is ~7.8dB and 9.1dB, respectively for ~5.5kpbs data rate.   
Source [14, xiaomi] observed that with up to 10% SFO, Option 2 of D2R preamble+3 midambles and Option 4 of D2R preamble+3 midambles+postamble can achieve 10% BLER when the SNR is within the range of [15, 25] dB for 1.25 kbps data rate. 
Source [30, Qualcomm] observed that the required SFO estimation accuracy to achieve 1% and 10% BLER is much smaller than 10^3 ppm. To achieve the required accuracy,
For Option 2 of D2R preamble+X midamble(s) where midamble inserted per every certain number of PDRCH bits (e.g., 192 bits),
For SFO estimation using each amble for the subsequent PDRCH bits (e.g., 192 bits), with up to 10% SFO, more than 50 SFO hypotheses are necessary at the reader side and with up to 1% SFO, 6 SFO hypotheses are sufficient at the reader side.
For SFO estimation based on the time gap between preamble and midamble, with up to 10% SFO, 10 SFO hypotheses are used, but reader has to store the received samples and wait for the midamble to start SFO/channel/interference estimation, demodulation, decoding, etc.
For Option 3 of D2R preamble+postamble, SFO estimation is based on the time gap between preamble and postamble, with up to 10% device SFO, 10 SFO hypotheses are used for reader, but reader has to store the received samples and wait for the postamble that is after the end of PDRCH for any of SFO/channel/interference estimation, demodulation, decoding, etc

For the synchronization and timing tracking of D2R transmission, 
Source [5, CMCC] report that with up to 10% SFO, option 1 is not sufficient for D2R reception since the residual SFO at reader side is larger than 1%. While with option 3, the reader can precisely search and detect the SFO with a residual SFO of 0.03% at -3dB SNR TDL-A channel.
Source [14, xiaomi] report that 
For packet size of 96bits, when the SNR is increased from -4dB to 20dB, the ratio of device residual SFO over 100ppm decreases to 6% for Option 2, 3 and 4, but remains at 95% for Option 1.
For packet size of 400bits, when the SNR is increased from -4dB to 20dB, the ratio of device residual SFO larger than 10ppm decreases to 5% for Option 2, 3, and 4, but is still 99.6% for Option 1.
Sources [9, vivo], [15, CATT] report that SFO estimation based on D2R preamble can achieve accurate estimation without additional ambles (midamble or postamble). 
Source [9, vivo][7 Samsung] observed that for non-coherent detection of PDRCH, the number of SFO hypotheses and the SNR needed for 10% and 1% BLER cannot significantly be reduced for option 2, 3 and 4 compared to the option 1. Moreover, the additional ambles i.e., midamble(s) and/or postamble introduces additional overhead and prevents pipelined processing of the reception.     
Source [15, CATT] observed that 
The coarse estimation of SFO based on the D2R preamble indicates that the SFO estimation error is less than 1% with a probability of 99.3%, and less than 0.1% with a probability of 49.9%.
The fine estimation of SFO based on the D2R preamble shows that the SFO estimation error is less than 1% with a probability of 99.5%, and less than 0.1% with a probability of 90.8%.
Reader/gNB can achieve a probability of not less than 99.5% for SFO estimation error below 1%, and 90.8% for SFO estimation error below 0.1% by receiving D2R preamble signals.
Source [30, Qualcomm] report that for D2R with coherent demodulation at reader, the reader needs to estimate the device clock frequency with the accuracy of 0.5% (5 * 10^3 ppm) or lower for a short message (e.g., 72 bits after CRC/coding) and 0.1% (10^3 ppm) or lower for a long message (e.g., 224 bits after CRC/coding). The source further reports that design of D2R amble(s) (e.g., overhead) and the correspondingly required number of SFO hypothesis for the estimation depend on the sampling clock accuracy that the device uses for D2R. 
Note: in the observations above where coherent detection is used, sources that evaluated option 3 and option 4 assumed that the postamble is used at least for time/frequency tracking and for channel estimation.


Agreement
For the CFO calibration signal, which is required only for device 2b to reduce the frequency offset range and the guard-bandwidth of D2R transmission, the following observations are captured in TR 38.769:
Source [3, Huawei] report that a single-tone RF signal is used as the CFO calibration signal, it is not a part of time acquisition signal and can be transmitted as an optional R2D signal after the PRDCH transmission. 
Sources [2, Ericsson], [19, Panasonic] and [20, OPPO] report that additional synchronization signal is needed. 
[OPPO] state the R2D timing acquisition signal may not be sufficient or may not be usable for CFO calibration since a reference frequency is needed when separate LOs are used for Tx and Rx in device 2b.
Sources [7, Samsung], [9, vivo], [30, Qualcomm], [36, Apple] report that additional synchronization signal is needed if the synchronization for carrier frequency using R2D signal/channel does not provide required functionalities for device 2b.
Source [5, CMCC][31, MTK] report that it may not be possible to achieve enough frequency accuracy (0.01 ppm) even after CFO calibration based on R2D time acquisition signals for coherent detection at reader especially when the D2R data rate is low.

Agreement
For device 2b, a signal for CFO calibration should be provided to synchronize / calibrate the device clock for LO for carrier frequency (Clock purpose #5) to achieve the accuracy after clock sync / calibration at device side captured in Table 5.2.3-1.
Frequency calibration at device 2b is beneficial at least to reduce the guard-bandwidth of D2R transmission.

Agreement
Adopt the updates documented in R1-2410653 for section 6.2 of the TR38.769. 

Agreement
Adopt following update to the TP agreed on Monday

Capture following observations in the TR 38.769, where CFO is assumed to be zero or negligible.
[omit unchanged part]
For coherent detection of PDRCH with a payload of 96bits with 16-bit CRC (or 6-bit CRC [14, Xiaomi]), using 1/2 Manchester coding and 1/3 or 1/2 convolutional code,
Sources [3, Huawei], [5, CMCC] and [14, xiaomi] observed that Option 1 cannot achieve 10% BLER.
Sources [6, ZTE], [7, Samsung], [9, vivo], [20, OPPO] and [30, QC] observed that Option 1 can achieve 10% BLER.
Sources [3, Huawei], [5, CMCC], [6, ZTE], [7, Samsung], [9, vivo], [14, xiaomi], [16, China Telecom] observed that adding additional amble improves the performance. 
Source [3, Huawei] observed that with up to 10% SFO, 
Option 2 of D2R preamble+ 1 midamble achieves 10% BLER at SNR around -3dB, but cannot achieve 1% BLER.
Option 3 of D2R preamble+ postamble achieves 10% BLER at SNR around -4dB, and can achieve 1% BLER at SNR around 4dB.
Source [5, CMCC] observed that with up to 10% SFO, Option 3 allows reader to precisely search and detect the SFO with 0.03% residual SFO at -3dB SNR TDL-A channel, achieving 10% BLER -2.44dB SNR for ~1 kbps data rate and -2.17 dB for ~6 kbps data rate. Source [5, CMCC] further observed that when the reader adopts same number of SFO hypothesis based on preamble, with 1% SFO, Option 3 can achieve 10% BLER at -4.27 dB SNR for ~1 kbps and at -4.29 dB SNR for ~6 kbps, which provides 1~2 dB performance gain when compared to 10% SFO. 
Source [14, xiaomi] observed that with up to 10% SFO, Option 2 of D2R preamble + 1 midamble, option 3, and option 4 of D2R preamble + 1 midamble+postamble achieve basically the same performance, the SNR for 10% BLER is 5dB for 1.25 kbps data rate.  
Source [6, ZTE] observed that with up to 10% SFO, ~1kbps data rate, and the same amble(s) overhead, Option 3 can provide 1~2 dB, 5dB performance gain for 10% BLER, 1% BLER, respectively, compared to Option 1. Additionally, Option 3 can provide ~1dB, 2dB performance gain for 10% BLER, 1% BLER, respectively, compared to Option 2 of D2R preamble +1 midamble.
Source [7, Samsung] observed that for ~5kbps data rate, compared to option 1, 
For device 1 with up to 10% SFO, Option 2 of D2R preamble + 1 midamble provides ~0.5 dB SNR gain at 10% BLER with TDL-A channel and ~0.9 dB SNR gain with TDL-D channel.
For device 2 with up to 1% SFO, Option 2 of D2R pramble + 1midamble provides ~1 dB SNR gain at 10% BLER with TDL-A channel and ~1.4 dB SNR gain with TDL-D channel.
Moreover, the additional ambles i.e., midamble(s) and/or postamble introduces additional overhead and postamble may prevent pipelined processing of the reception. 
Source [9, vivo] observed that, 
With up to 10% SFO, ~5.5kbps data rate, Option 1 achieves 10% and 1% BLER at SNR ~0.7dB and 10dB, respectively; Additionally, maintaining the same amble overhead, Option 2 (D2R preamble + 1 midamble) and Option 3 demonstrate similar performance, achieving 10% and 1% BLER at SNR around -1.7dB and 5.2dB, respectively.
With up to 1% SFO, ~5.5kbps data rate, Option 1 achieves 10% and 1% BLER at SNR ~ -1.3dB and 11dB, respectively. Additionally, with the same amble overhead, the SNR difference between Option 2 (D2R preamble+1midamble) and Option 3 for 10% and 1% BLER is less than 1dB, with SNRs ~ -3.1dB to -2.5dB for 10% BLER and ~3.6dB to 4.5dB for 1% BLER.
Source [16, China Telecom] observed that with up to 10% SFO, ~7.5kbps data rate, there is ~6~7dB performance gap at 10% BLER and ~10.5~11.5dB performance gap at 1% BLER between option 2 of D2R preamble+111 midambles and option 1. Note that Source [16, China Telecom] does not use any convolutional code.

Source [20, OPPO] observed that with perfect SFO estimation, 1kbps data rate and OOK modulation, there is no noticeable performance gain from using midamble(s) and/or postamble for PDRCH transmission, regardless of the payload size.
Source [30, Qualcomm] observed that the required SFO estimation accuracy to achieve 1% and 10% BLER is <10^3 ppm. To achieve the required accuracy,
For Option 1, more than 50 SFO hypotheses at reader side are necessary for device with up to 10% SFO and 6 SFO hypotheses are sufficient at reader side for device with up to 1% SFO. 
For Option 3, 10 SFO hypotheses are sufficient for device with up to 10% SFO, but reader has to store the received samples and wait for the postamble that is after the end of PDRCH for any of SFO/channel/interference estimation, demodulation, decoding, etc.

For coherent detection of PDRCH with a payload of 400bits with 16-bit CRC, using 1/2 Manchester coding and 1/3 or 1/2 convolutional code, 
For option 1 of D2R preamble only, 
Sources [3, Huawei], [5, CMCC], [6, ZTE], [8, Spreadtrum], [9, vivo], [14, xiaomi] observed that with up to 10% SFO, 10% BLER cannot be achieved. 
Source [20, OPPO] observed that with perfect SFO estimation, 1kbps data rate and OOK modulation, there is no noticeable performance gain from using midamble(s) and/or postamble for PDRCH transmission.
For other amble options, 
Source [3, Huawei] observed that
With accurate SFO estimation, Option 2 of D2R preamble + 4 midambles can achieve 10% BLER at SNR ~ 2.7dB but cannot achieve 1% BLER.
With up to 10% SFO, Option 3 cannot achieve 10% BLER.
With up to 10% SFO, Option 4 of D2R preamble+2 midambles+postamble achieves 10% BLER at SNR of ~0.25dB; But it cannot achieve 1% BLER. Option 4 of D2R preamble+3 or 4 midambles+postamble, achieves a 10% BLER at an SNR of around -0.2 dB, and achieves 1% BLER at SNR around 9dB or 8dB, respectively.
Source [5, CMCC] observed that with up to 10% SFO, Option 4 of D2R preamble combined with 1 to 4 midambles + postamble, achieves 10% BLER at SNR of 2.5 dB, 1 dB, 0.8 dB, or 0.5 dB, respectively, for a data rate of around 1 kbps.
Source [6, ZTE] observed that with up to 10% SFO, 
Option 3 can provide ~5.5 dB performance gain compared to option 2 of D2R preamble+1midamble for 10% BLER, with the same amble(s) overhead for ~1kbps data rate.
Option 2 of D2R preamble+1midamble cannot achieve 1% BLER for ~1kbps data rate.
Option 4 of the D2R preamble+1 or 2 midamble(s)+postamble, has similar performance, it can achieve a 10% BLER at SNR of -1dB and achieves a 1% BLER at SNR of 6dB and 5dB respectively for ~1kbps data rate.  
Source [8, Spreadtrum] observed that with up to 10% SFO, 
Option 3 of D2R preamble+ postamble cannot achieve 10% BLER for ~7kpbs.  
Option 4 of D2R preamble + 1 midamble + postamble can achieve 10% BLER and 1% BLER at SNR around -6dB and 0 dB, respectively for ~7kpbs data rate.
Source [9, vivo] observed that 
With up to 10% SFO, maintaining the same amble overhead, both Option 2 of D2R preamble+5 midambles and Option 4 of D2R preamble+4 midambles+1 postamble demonstrate similar performance, achieving 10% BLER at SNR ~0.2dB. While for 1% BLER, the SNR for Option 2 and Option 4 is ~ 9.2dB and 12.8dB, respectively for ~5.5kpbs data rate.   
With up to 1% SFO, maintaining the same amble overhead, both Option 2 of D2R preamble+5 midambles and Option 4 of D2R preamble+4 midambles+1postamble demonstrate similar performance, achieving 10% BLER at SNR around -1.2dB. While for 1% BLER, the SNR for Option 2 and Option 4 is ~7.8dB and 9.1dB, respectively for ~5.5kpbs data rate.   
Source [14, xiaomi] observed that with up to 10% SFO, Option 2 of D2R preamble+3 midambles and Option 4 of D2R preamble+3 midambles+postamble can achieve 10% BLER when the SNR is within the range of [15, 25] dB for 1.25 kbps data rate. 
Source [30, Qualcomm] observed that the required SFO estimation accuracy to achieve 1% and 10% BLER is much smaller than 10^3 ppm. To achieve the required accuracy,
For Option 2 of D2R preamble+X midamble(s) where midamble inserted per every certain number of PDRCH bits (e.g., 192 bits),
For SFO estimation using each amble for the subsequent PDRCH bits (e.g., 192 bits), with up to 10% SFO, more than 50 SFO hypotheses are necessary at the reader side and with up to 1% SFO, 6 SFO hypotheses are sufficient at the reader side.
For SFO estimation based on the time gap between preamble and midamble, with up to 10% SFO, 10 SFO hypotheses are used, but reader has to store the received samples and wait for the midamble to start SFO/channel/interference estimation, demodulation, decoding, etc.
For Option 3 of D2R preamble+postamble, SFO estimation is based on the time gap between preamble and postamble, with up to 10% device SFO, 10 SFO hypotheses are used for reader, but reader has to store the received samples and wait for the postamble that is after the end of PDRCH for any of SFO/channel/interference estimation, demodulation, decoding, etc
Source [7, Samsung] observes that the additional ambles i.e., midamble(s) and/or postamble introduces additional overhead and postamble may prevent pipelined processing of the reception.

For the synchronization and timing tracking of D2R transmission, 
Source [5, CMCC] report that with up to 10% SFO, option 1 is not sufficient for D2R reception since the residual SFO at reader side is larger than 1%. While with option 3, the reader can precisely search and detect the SFO with a residual SFO of 0.03% at -3dB SNR TDL-A channel.
Source [14, xiaomi] report that 
For packet size of 96bits, when the SNR is increased from -4dB to 20dB, the ratio of device residual SFO over 100ppm decreases to 6% for Option 2, 3 and 4, but remains at 95% for Option 1.
For packet size of 400bits, when the SNR is increased from -4dB to 20dB, the ratio of device residual SFO larger than 10ppm decreases to 5% for Option 2, 3, and 4, but is still 99.6% for Option 1.
Sources [9, vivo], [15, CATT] report that SFO estimation based on D2R preamble can achieve accurate estimation without additional ambles (midamble or postamble). 
Source [9, vivo][7 Samsung] observed that for non-coherent detection of PDRCH, the number of SFO hypotheses and the SNR needed for 10% and 1% BLER cannot significantly be reduced for option 2, 3 and 4 compared to the option 1. Moreover, the additional ambles i.e., midamble(s) and/or postamble introduces additional overhead and postamble may prevents pipelined processing of the reception.     
Source [15, CATT] observed that 
The coarse estimation of SFO based on the D2R preamble indicates that the SFO estimation error is less than 1% with a probability of 99.3%, and less than 0.1% with a probability of 49.9%.
The fine estimation of SFO based on the D2R preamble shows that the SFO estimation error is less than 1% with a probability of 99.5%, and less than 0.1% with a probability of 90.8%.
Reader/gNB can achieve a probability of not less than 99.5% for SFO estimation error below 1%, and 90.8% for SFO estimation error below 0.1% by receiving D2R preamble signals.
Source [30, Qualcomm] report that for D2R with coherent demodulation at reader, the reader needs to estimate the device clock frequency with the accuracy of 0.5% (5 * 10^3 ppm) or lower for a short message (e.g., 72 bits after CRC/coding) and 0.1% (10^3 ppm) or lower for a long message (e.g., 224 bits after CRC/coding). The source further reports that design of D2R amble(s) (e.g., overhead) and the correspondingly required number of SFO hypothesis for the estimation depend on the sampling clock accuracy that the device uses for D2R. 
Source [37, MediaTek] reports that transmitting 96-bit packet size with 16-bit CRC requires residue SFO after reader compensation to be 1000 ppm, and transmitting 1000-bit packet size with 16-bit CRC requires residue SFO after reader compensation to be 100 ppm.
Note: in the observations above where coherent detection is used, sources that evaluated option 3 and option 4 assumed that the postamble is used at least for time/frequency tracking and for channel estimation.

Agreement
Following observations on R2D clock-acquisition part are captured in TR 38.769:
On impact/restriction of M values for the clock-acquisition part
9 sources [TCL, Nokia, Huawei, CMCC, ZTE, Apple, CATT, Mediatek, Qualcomm] provided observations on the impact/restriction of M values for the clock-acquisition part design requirements:
1 source [Nokia] observed that increasing value of M, while retaining the same transmission duration, improves the auto-/cross- correlation properties of the sequence due to increase in sequence length and use of   provides better timing estimation accuracy even in the presence of SFO as the sequence length spans only over a shorter duration. 
2 sources [TCL, Huawei] observed for option 1 of the clock-acquisition part design that no restriction is required to be placed on the M values. Furthermore, 1 source [Huawei] observed that the same 2 ON-OFF voltage (with the same duration) satisfies the FDR performance metric of less than 1% for different M values, e.g., M = 2, 6 and 24, where FDR is the False detection ratio (FDR), i.e. incorrectly calculating M, is the performance metric.
1 source [CMCC] observed that pattern of the clock-acquisition part is related to M chips per OFDM symbol and when M is small, the clock-acquisition part may cross multiple OFDM symbols, and the CP insertion may degrade the timing acquisition performance.
1 source [ZTE] observed that with option 2, the duration of the clock-acquisition part remains consistent across all M values, at least three OFDM symbols maybe required for clock-acquisition part and it maybe not as efficient as option1
1 source [Apple] observed that among the two options studied for the clock-acquisition part, option 2 provides increased robustness, especially in case of large value of M, when compared to option 1 and potentially increase the detection performance of the clock-acquisition part. 
1 source [CATT] observed that if the chip duration is variable based on the M value used for OOK-4 waveform, the detection performance would be limited by the received SINR of the CAP with clear transition of the rising and falling edges.
1 source [Qualcomm] observed that the option 1 with M>1 has shorter duration of clock acquisition part than M=1 and worse timing acquisition accuracy. At least part of PRDCH following the clock acquisition part may need to be used to improve the timing acquisition. Furthermore, the larger M (e.g., M>4) with small chip duration is more sensitive to the SFO accuracy and the restriction of M for the clock acquisition part may be needed.
1 source [Mediatek] further observed that different M values may impact the chip accuracy obtained by the clock acquisition part.

On impact of CP insertion/handling on the clock-acquisition part
10 sources [TCL, CMCC, ZTE, Samsung, Vivo, CATT, NTT Docomo, Qualcomm, Mediatek, Spreadtrum] observed that the CP insertion/handling may impact the design requirements of the clock-acquisition part:
1 source [CMCC] further observed that when the clock-acquisition part occupies more than one OFDM symbol, ON-OFF state transition around CP can avoid the error rising or falling edges due to the CP insertion.
1 source [ZTE] further observed that to mitigate the impact of the CP in the clock-acquisition part for large M values, it can reuse the CP handling method for PRDCH 
1 source [Samsung] further observed that CP insertion/handling on the clock-acquisition part can cause false rising/falling transition and, therefore, the clock acquisition part should be designed such that it does not incur a false rising or falling edges due to CP insertion when CP-OFDM is used for OOK signal generation.
1 source [vivo] further observed that CP insertion/handling on the clock acquisition part will impact the chip duration estimation accuracy. It is further observed that for CP handling, device may not be able to count the clock and estimate OFDM symbol duration accurately until the clock acquisition part if the start indicator only includes a single ON-OFF transmission. 
1 source [CATT] further observed that the SER will be degraded due to uneven chip interval when the CP is inserted within an OFDM symbol, where SER refers to the number of samples which is mismatched for comparing to the total number of samples in a chip.
1 source [NTT Docomo] further observed if CP insertion would cause false rising/falling edges, accuracy of timing acquisition may be impacted.   
1 source [Mediatek] further observed that the issues of chip extension, false raising/falling transition, and additional raising/falling transition caused by CP insertion/handling considering different M values will impact the chip accuracy obtained by the clock acquisition part.
1 source [Spreadtrum] further observed that the design of clock acquisition part should consider that CP insertion does not cause a false rising or falling edges and does not cause different length of multiple high / low voltages within the clock acquisition part when the clock acquisition spans multiple OFDM symbols.
1 source [Huawei] observed CP insertion/handling may not impact the design requirements of the clock-acquisition part

Agreement
For the D2R preamble design, following aspects have been studied and can be captured in the TR 38.769:
Autocorrelation Property
10 sources [Nokia, Huawei, CMCC, Xiaomi, CATT, Oppo, Ericsson, NTT Docomo, Qualcomm, ZTE] observed that the signal should have good autocorrelation properties for accurate peak detection based on the signal correlation at the reader 
Cross-correlation Property
7 sources [Nokia, CMCC, Oppo, Ericsson, Qualcomm, ZTE, CATT] observed that the signal should have good cross-correlation properties if multiple D2R preamble sequences are considered (e.g. for multiple access schemes (if supported) for D2R transmissions). 
Line coding
1 source [Nokia] observed that line coding may impact the autocorrelation property of the sequence. 
1 source [Huawei] observed that for D2R preamble, to apply backscattering, line coding can help improve the detection performance based on shifting the D2R signal’s frequency location away from the carrier wave
Sequence Types (not limited to below types only)
M-sequence
3 sources [Nokia, Vivo, Xiaomi] observed that m-sequence can be considered for D2R preamble mainly owing to good correlation properties.  
Golay sequence
4 sources [CMCC, Vivo, Xiaomi, Samsung] observed that Golay sequence can be considered for D2R preamble mainly owing to good correlation properties and availability of large number of distinct sequences and complementary pairs.  
Walsh sequence
1 source [Oppo] observed that Walsh sequence can be considered as a candidate for D2R preamble thanks to its good auto/cross-correlation property and flexible length
General Observations
1 source [Huawei] observed can achieve 0.97% residual SFO with 98% probability under -2.5dB SNR and 0.1% MDR with [-1/8, 1/8] chip timing error with 99.05% probability under -2.5dB SNR with D2R preamble including 2-parts with clock-like sampling frequency signal and timing-acquisition signal, having 32-length ‘1’ sequence (encoded to 64-chip Manchester code) and 32-length sequence (encoded to 64-chip Manchester code), respectively.
4 sources [TCL, CMCC, ZTE, Vivo] observed that for D2R preamble with binary signal, the timing synchronization performance is highly related to the sequence length of the preamble. Furthermore, 1 source [CMCC] observed that to achieve a BLER performance at 10%, the timing synchronization error should be less than 10%. Furthermore, 1 source [ZTE] observed that the channel estimation performance is also highly related to sequence length. 1 source [ZTE] observed that using a 32 bits preamble provides ~8 dB, ~5 dB performance gain than using 8 bits, 16 bits preamble, respectively. And using a 64 bits preamble provides ~2.5dB performance gain than using a 32 bits preamble.
1 source [Ericsson] observed that for D2R preamble with binary signal, normalized SFO estimation error of less than 10% can be achieved with a training sequence length 64 or longer. The simulated D2R preamble consisting of a Golay complementary pair can tolerate SFO up to 1% (AWGN) with up to 1 dB loss in performance for a sufficiently long preamble sequence length (32 or greater).

Agreement
For determining the end of PRDCH at the device, following two options are studied and captured in the TR 38.769:
Option 1: TBS information (via implicit/explicit L1 R2D control information)
Option 2: Postamble (at the end of PRDCH) 
14 sources [Nokia, Huawei, ZTE, CMCC, Samsung, Ericsson, Oppo, LGE, Qualcomm, Spreadtrum, Mediatek, Cewit, Ericsson, vivo] provided following observations on the above two options for determining the end of PRDCH:
3 sources [Nokia, Huawei, ZTE] observed that option 2 provide two benefits, namely, the variable payload length and to provide timing acquisition before the subsequent transmission of either PDRCH or PRDCH, thus improving the detectability at both reader and the device, respectively. Furthermore, 1 source [Huawei] observed that R2D postamble indicates the TBS with high efficiency for small packets by avoiding a large padding overhead, unlike option 1, which may require devices to perform blind detection of different PRDCH formats (if supported) and the overhead caused by the inclusion of a R2D postamble does not exceed 20% for even the smallest of message sizes and may be less than the signaling overhead caused by using a dedicated TBS indicator
1 source [CMCC] observed for option 2, that for small payload size with only a few bits, the presence of long postamble generates large resource overhead, while for large payload size with more bits, the resource overhead of postamble is smaller.
1 source [vivo] observed for option 2, that for small payload size with only a few bits, the presence of long postamble generates large resource overhead.
1 source [Samsung] observed option 2 is not strictly required, however, given the possible clock drift at a device, it may be still beneficial to also attach postamble at least for the determination of the end of PRDCH at a device. 
3 sources [Oppo, Spreadtrum, CEWiT] observed that with option 2, the false detection may be higher for shorter postamble. Source [OPPO[ observed that in contrast to option 2, it is more reliable and efficient to indicate TBS with control information in option 1
2 sources [LGE, vivo] observed that if a message type or a command ID is included in L1 control information and implicitly indicates a known size of a fixed TB, then there is no need for either option 1 or option 2
2 sources [Qualcomm, vivo] observed that option 1 has the advantages of avoiding blind detection of postamble and providing the power saving for non-target devices to skip the R2D detection.
1 source [MediaTek] observed that option 1 is feasible for the device to avoid the unnecessary reception of a TB with a specific size and thus enable power saving, e.g., when the TB has a size exceeding the allowance of the device remaining power.
1 source [Ericsson] observed option 2 is not strictly required if the end of PRDCH can be explicitly indicated by R2D control information, and it is subject to the miss-detection rate. It may be beneficial if a PRDCH postamble can serve as an additional timing acquisition signal prior to a PDRCH transmission.

Agreement
For D2R scheduling, midamble (if supported) related information can be explicitly/implicitly indicated via corresponding PRDCH.

Agreement
Following observations on R2D clock-acquisition part are additionally captured in TR 38.769:
On purpose of SFO estimation/correction based on the clock-acquisition part
3 sources [Nokia, CATT, Qualcomm] provided observations on the applicability of clock-acquisition part for frequency synchronization:
1 source [Nokia] observed that the length of preamble sequence may need to consider also the robustness against SFO
1 source [CATT] observed that device 2a/2b may require higher synchronization accuracy for signal transmission or backscattering and therefore, the design of CAP may be required to accommodate the requirement of additional frequency synchronization and clock calibration for Device 2a/2b. 
1 source [Qualcomm] further observed for Option 1, as the CAP duration with high M is decreased, only CAP may not be sufficient for SFO correction and for Option 2, as the CAP duration is fixed and independent from M, the CAP with long enough duration can support SFO correction.
On purpose of CFO estimation/correction based on the clock-acquisition part
2 sources [Ericsson, Qualcomm] provided observations on the applicability of clock-acquisition part for frequency synchronization:
1 source [Ericsson] observes that the clock-acquisition part can be utilized to solve the frequency synchronization problem without impacting the time-domain sequence, for example by transmitting in some frequency resources and it can be a harmonized solution for both chip duration indication and device frequency synchronization. However, it is further observed that if the time interval between an R2D transmission and the corresponding D2R transmission and if the device loses the timing obtained from the R2D timing acquisition signal due to timing drift at the time for the D2R transmission, then an additional synchronization signal is needed
1 source [Qualcomm] further observed for Option 1, as the CAP duration with high M is decreased, only CAP may not be sufficient for CFO correction and for Option 2, as the CAP duration is fixed and independent from M, the CAP with long enough duration can support CFO correction.


WI Phase: RAN1 Agreements (relevant for R2D/D2R signals including timing acquisition and synchronization
RAN1#120 (Athens, Greece, Feb 17th – 21st, 2025)
SIP related Agreements
Agreement
For the SIP of R-TAS, for providing the start of the R2D transmission, one single design based on Option 1 is supported and further down-selection to be done among Alt 1 and Alt 2 :
Option 1: ON-OFF transmission with following alternatives:
Alt 1: A single ON-OFF transmission with pre-defined duration for each of the ON-OFF, where ON and OFF may have same or different durations
Continue discussion to clarify the duration of each of the ON and OFF
Continue discussion to list the different candidate proposals under Alt1
Alt 2: A multi-ON-OFF transmission with pre-defined duration for each of the ON(s)-OFF(s), where different ON and different OFF may have same or different durations and different parts may have same or different duration
Continue discussion to clarify the duration of each of the ON and OFF
Continue discussion to list the different candidate proposals under Alt2
Only a single fixed value for entire duration of SIP of R-TAS is supported, which is independent of the value of “M” used in CAP and PRDCH
Note: Specific design and duration for SIP of R-TAS are further discussed, and companies are encouraged to evaluate the designs in terms of target MDR of [10%] for a FAR up to [1%] and at least following assumptions are used:
MDR refers to the probability that SIP is not detected when it was actually transmitted
FAR probability that the receiver incorrectly detects SIP when SIP was not transmitted
Energy/edge detection-based method is the baseline assumption for evaluation purpose
Continue discussion on necessary details for simulation assumptions

Agreement
For the SIP of R-TAS, down-select among the following candidates:
Alt 1 (Single ON-OFF transmission)
Alt 1-1: ON followed by OFF with same duration for both
Alt 1-2: ON followed by OFF with a duration ratio of 1:[2,3]
Alt 1-3: ON followed by OFF with a duration ratio of [2,3]:1
Alt 2 (Multi-ON-OFF transmission)
Alt 2-1: A number of repetition instances of Alt 1-1 or Alt 1-2 or Alt 1-3
Alt 2-2: ON-OFF-ON (duration of ON and OFF can be different)
Alt 2-3: OFF-ON-OFF (duration of ON and OFF can be different)
Alt 2-4: Combination of single instance of Alt 1-1 and single instance of Alt 1-2
For the evaluation purpose, for both options, candidate values related to duration are considered:
Entire duration of SIP: 1/2 OFDM symbol duration or 1 OFDM symbol duration (including clarifying whether OFDM symbol duration includes CP); additional durations can be considered and reported by companies with justification 
Companies to report the exact duration(s) for ON or OFF
Companies are encouraged to report at least the following details for the evaluations:
Baseline assumption is that RF transmission is not present; companies can report other consideration
For FAR calculation, whether noise and/or PRDCH transmission is considered
Details on threshold detection method including whether/how threshold detection training is used based on the proposed design alternative or not
BW assumptions for RF-ED and BB-LPF
Target MDR of up to 1% for FAR of up to [1%, 10%]

CAP related Agreements
Agreement
For the CAP of R-TAS, the starting chip has a different voltage level compared to the end of the SIP of R-TAS.

Agreement
For the design of the CAP of R-TAS, at least 2 transition edges in same direction are included, i.e. at least two transitions from “OFF” chip to “ON” chip or two transitions from “ON” chip to “OFF” chip.

Agreement
For the CAP of R-TAS:
Candidate values for maximum duration of CAP to be further down-selected to one value from : 1.5 OFDM symbol duration, 2 OFDM symbol duration, 3 OFDM symbol duration
For option 1 for CAP of R-TAS from TR 38.769, maximum duration is applicable to minimum value of M to be supported, and the CAP duration becomes shorter with increasing value of M
FFS: whether the number of ON/OFF transmissions in the CAP is fixed or not fixed
For option 2 for CAP of R-TAS from TR 38.769, maximum duration is the only (constant) duration that is applicable for all the M values to be supported
Down-selection between option 1 and option 2 for CAP of R-TAS from TR 38.769 by RAN1#120-bis
FFS: Values of M to be supported


R2D Midamble related Agreement
Agreement
R2D transmission does not include a midamble.

D2R X-amables related Agreement
Agreement
For D2R preamble design, the functionalities of timing acquisition, SFO estimation/time tracking and channel estimation should be supported
For D2R midamble design, the functionalities of SFO estimation/time tracking and channel estimation should be supported
D2R midamble can be transmitted at the end of the PDRCH transmission. If it is at the end, it is not designed for being used for indicating the end of PDRCH transmission
FFS: condition(s) and/or indication where the D2R midamble is present or not

Agreement
For D2R x-ambles:
Following is considered as the types for base sequence and to be further down-selected:
Option 1: M-sequence 
Option 2: Golay sequence
Note: Above doesn’t preclude an additional part for preamble, e.g. with ON and/or OFF transmission, if needed/supported
FFS: Whether/what multiple sequences (using same base sequence type) are supported
Note: This in no way implies that there is going to be CDMA between D2R x-ambles
For evaluation purpose, companies are encouraged to consider following:
Performance at least in terms of autocorrelation/cross-correlation property, SFO estimation/Timing accuracy, SNR for target PDRCH BLER of [1%, 10%]
Report presence and time-domain resource(s) x-ambles
Report sequence type(s) and length(s) for x-ambles
Following format can be considered for reporting the evaluation results


RAN1#120bis (Wuhan, China, April 7th – 11th, 2025)
Agreement
For D2R midamble, for determining the presence and location of midamble(s) at the device:
Reader explicitly indicates the same interval between consecutive midambles, and between the preamble and the first midamble, via R2D control information
FFS: details of signalling
FFS: whether the reader can explicitly indicate with one bit whether a midamble is additionally present at the end
Note: This does not preclude the support of having no midamble present in the D2R transmission

Agreement
For the pattern of SIP of R-TAS, only the following 2 alternatives are considered for further down-selection:
Alt 1-2: ON-OFF with a ratio of 1:3 and with following total SIP duration to be further down-selected:
Option 1: 0.5 OFDM symbol duration
Option 2: 1 OFDM symbol duration
Alt 2-4: ON-OFF-ON-OFF with a ratio of 1:1:1:3 and with following total SIP duration to be further down-selected:
Option 1: 0.5 OFDM symbol duration
Option 2: 1 OFDM symbol duration

TDoc file conclusion not found

3GPP TSG RAN WG1 #120bis		        					     R1-2502611
Wuhan, China, April 7th – 11th, 2025

Agenda Item:	9.4.3
Source: 	Moderator (Apple)
Title:	FL Summary#3 on timing acquisition & synchronization for Ambient IoT
Document for:	Discussion & Decision

Introduction & Work-Plan for RAN1#120bis
This document provides the feature lead summary on the offline discussions/inputs/proposals for AI 9.4.3 timing acquisition and synchronization signals for R2D and D2R for ambient IoT WI during RAN1#120bis. 


Contact Information
Please consider providing your company name, your name and email address to be able to reach for any potential offline discussions/contact regarding AI 9.4.3 on timing acquisition and synchronization for ambient IoT.





Work Plan
Following is my plan for this meeting for this agenda for RAN1#120bis:

SIP of R-TAS
For SIP, at least down-select to not more than 2 alternatives from the agreed alternatives in RAN1#120 considering technical justification and evaluation analysis. Also agree on fixed single duration for the down-selected alternatives. 
CAP of R-TAS
For CAP, plan is to down-select between option 1 and option 2 (as agreed in last meeting) and then also limit the number of specific CAP pattern design and hopefully have not more than 2 patterns for further down-selection by end of this meeting
R2D Postamble aspects
For indicating the end of PRDCH transmission, at least down-select to whether R2D control information is used, or Manchester coding rule violation is used. 
D2R X-amble
For D2R x-amble, first priority is to agree on the sequence type and length(s). In addition, for midambles, agree on the method/signaling that allows the device to determine the number of midamble(s) and their location within the PDRCH.

SIP of R-TAS

[Closed] 1st Discussion Round 
For the SIP of R-TAS, following aspects have been discussed in the contributions following the agreements and discussion from RAN1#120:
Preferred alternative(s) for SIP design
Duration of SIP
For comparison among different alternatives, companies also provided simulations results/analysis to justify their preferred alternative(s). In addition, a few companies also discussed other aspects including M value/chip duration for the SIP, whether/what impact of CP insertion on SIP. In the following table, companies’ views on their preferred alternatives along with justification and performance analysis, if provided are summarized. 

Table 2-1: Summary of views on alternatives for SIP



FL Observations:
28 companies provided their views on SIP of R-TAS
11 companies provided evaluation in terms of target MDR and FAR with most of the companies considering MDR <= 1% for a target FAR <= 1%
For Alt 1-1, 3 companies provided their preference with ON-OFF ratio of 1:1 & total duration spanning 1 OFDM symbol and only 1 company [China Telecom] showed that it performs better than other alternatives while it satisfies the target MDR and FAR. On the other hand, 4 companies [Futurewei, Xiaomi, Samsung, Huawei] showed other alternatives that perform better than Alt 1-1.
For Alt 1-2, overall, 8 companies provided their preference for at least some variant of Alt 1-2 and 6 companies [Futurewei, ZTE, Spreadtrum, Oppo, Xiaomi, Qualcomm] showed that at least some variant of Alt 1-2 performs better than other alternatives including Alt 1-1, Alt 1-3, Alt 2-2, Alt 2-4. On the other hand, 4 companies [China Telecom, Samsung, CMCC, Huawei] showed other alternatives that performs better than Alt 1-2. Within Alt 1-2:
2 companies prefer ON-OFF with ratio of 1:3 & total duration spanning 0.5 OFDM symbol
5 companies prefer ON-OFF with ratio of 1:3 & total duration spanning 1 OFDM symbol
3 companies prefer ON-OFF with ratio of 1:2 & total duration spanning 1 OFDM symbol
For Alt 1-3, overall, 6 companies provided their preference for at least some variant of Alt 1-3 and only 1 company [Samsung] showed that at least some variant of Alt 1-3 performs better than other alternatives including Alt 1-1, Alt 1-2. On the other hand, 6 companies [China Telecom, ZTE, Futurewei, Oppo, Qualcomm, Huawei] showed other alternatives that performs better than Alt 1-3. Within Alt 1-3:
4 companies prefer ON-OFF with ratio of 2:1 & total duration spanning 1 OFDM symbol
5 companies prefer ON-OFF with ratio of 3:1 & total duration spanning 1 OFDM symbol
For Alt 2-1, 7 companies provided their preference with ON-OFF ratio of 1:1 & total duration spanning 1 OFDM symbol and no company showed that it performs better than other alternatives. On the other hand, 2 companies [China Telecom, Vivo] showed other alternatives that perform better than Alt 2-1.
For Alt 2-2, overall, 6 companies provided their preference for at least some variant of Alt 2-2 and only 1 company [Vivo] showed that at least some variant of Alt 2-2 performs better than other alternatives including Alt 2-1, Alt 2-4. On the other hand, 2 companies [China Telecom, ZTE] showed other alternatives that performs better than Alt 2-2. Within Alt 2-2:
1 company prefer ON-OFF-ON with ratio of 1:1:4 & total duration spanning 1 OFDM symbol
2 companies prefer ON-OFF-ON with ratio of 1:4:1 & total duration spanning 1 OFDM symbol
3 companies prefer total duration spanning 1 OFDM symbol without any specific consideration on ration of ON/OFF
For Alt 2-3, overall, 3 companies provided their preference for at least some variant of Alt 2-3 and no company showed that it performs better than other alternatives including. On the other hand, 2 companies [China Telecom, Huawei] showed other alternatives that performs better than Alt 2-3. Within Alt 2-3:
2 companies prefer OFF-ON-OFF with ratio of 1:1:4 & total duration spanning 1 OFDM symbol
1 company prefer OFF-ON-OFF with ratio of 1:4:1 & total duration spanning 1 OFDM symbol
1 company prefer OFF-ON-OFF with total duration spanning 1 OFDM symbol
For Alt 2-4, 4 companies provided their preference with ON-OFF-ON-OFF ratio of 1:1:1:3 & total duration spanning 0.5 OFDM symbol and 2 companies [CMCC, Huawei] showed that it performs better than other alternatives. On the other hand, 2 companies [ZTE, Vivo] showed other alternatives that perform better than Alt 2-4.
A few companies discussed M value for SIP and whether/how the selection of M may impact the SIP detection considering CP insertion. However, based on discussions during RAN1#120, we considered discussing in terms of OFDM symbol duration and avoid reference M value. Therefore, with that understating, we should not try to reopen the discussion of reference M value but rather have discussion in terms of OFDM symbol duration. Eventually, it will be up to the specification editor on how to capture the SIP duration and corresponding ON(s) and OFF(s). Furthermore, at least 1 company raised the issue that for chip duration corresponding to M < 6, minimum PRB of 2 will be needed. However, another company clarified that we don’t need to associate the mapping of M to minimum number of PRBs for SIP and device should be able to adjust its LPF between SIP reception and follow-up R2D reception. 
Another aspect that a few discussed is whether/what impact CP might have on SIP detection. One company pointed out that chip duration corresponding to higher value of M map impact SIP detection, while 3 companies pointed out that CP impact is negligible on SIP and should not be considered for specific optimization, especially considering that all companies considered SIP duration of not more than 1 OFDM symbol

Based on above observations from companies’ contributions, as a starting point, FL proposal would be to down-select to Alt 1-2 and Alt 2-4. Main rationale behind these two alternatives is not just majority preference but evaluations showing that these two alternatives at least perform better within alternative set 1 and set 2, respectively. Further focused discussion on Alt 1-2 and Alt 2-4 can be further done, if we down-select these 2 options. Therefore, accordingly Proposal 2-1 is provided below:

(Updated) Proposal 2-1 
For the pattern of SIP of R-TAS, only following 2 alternatives are considered for further down-selection to one alternative:
Alt 1-2: ON-OFF with a ratio of 1:3 and with following total SIP duration to be further down-selected [during RAN1#120bis]:
Option 1: 0.5 OFDM symbol duration
Option 2: 1 OFDM symbol duration
Alt 2-4: ON-OFF-ON-OFF with a ratio of 1:1:1:3 and with following total SIP duration to be further down-selected [during RAN1#120bis]:
Option 1: 0.5 OFDM symbol duration
Option 2: 1 OFDM symbol duration





CAP of R-TAS


[Closed] 1st Discussion Round 
For the CAP of R-TAS, following aspects have been discussed in the contributions following the agreements and discussion from RAN1#120:
Preferred option in terms of variable or fixed duration for CAP design
Preferred exact CAP pattern including M values/duration and CP insertion related aspects

Table 3-1: Summary of views between option 1 and option 2 for CAP




Table 3-2: Summary of views on exact CAP pattern


FL observations
Overall, 30 companies provided their views on CAP of R-TAS
26 companies provided their preference relation to option 1 and option 2 from the TR and 17 companies indicated their preference for option 1, while 8 companies indicated their preference for option 2. 1 company indicated the preference to support and combine both options. Based on the agreement from RAN1#120bis, one option needs to be down-selected in this meeting. From the proponents of option 1, 3 companies provided evaluations and showed that option 1 is able to achieve target FDR, as long as there is sufficient gap between different M values. On the other hand, from the proponents of option 2, 2 companies provided evaluations and showed that at least for higher M values, option 2 is needed. 
Among specific CAP patterns, 5 alternatives are provided. Among the 5 alternatives, majority shows that either 3 chip with alternate ON/OFF or 4 chips with alternate ON/OFF can work well for the M values. Alt 3, 4, 5 with more chips have been proposed only by 1-2 companies each and don’t offer any significant gain compared to Alt 1 and Alt 2 and rather increase the overhead, especially for lower M values
Based on above, proposal 3-1 and proposal 3-2 are provided.



(Updated) Proposal 3-1
For CAP of R-TAS, option 1 for CAP of R-TAS from TR 38.769 is adopted with following details:
CAP duration becomes shorter with increasing value of M
FFS: exact relation between duration of CAP and M values



Proposal 3-2 
For the pattern of CAP of R-TAS, only following 2 alternatives are considered for further down-selection to one alternative:
Alt 1: 3 chips with ON-OFF-ON pattern 
Alt 2: 4 chips with ON-OFF-ON-OFF pattern
Based on the minimum value of M that will be agreed in agenda 9.4.2, corresponding maximum duration of CAP will be selected from the candidate set of {1.5 OFDM symbols, 2 OFDM symbols, 3 OFDM symbols}


[Closed] 2nd Discussion Round 
[Closed] Proposal 3-3
For CAP of R-TAS, following is adopted:
ON-OFF-ON-OFFON pattern with Option 1 for CAP of R-TAS from TR 38.769 is adopted where the CAP duration becomes proportionally shorter with increasing value of M, i.e. if for , duration is  OFDM symbol long, then for , duration is  OFDM symbol long


Proposal 3-3b
For CAP of R-TAS, following is adopted:
ON-OFF-ON-OFFON pattern with Option 1 for CAP of R-TAS from TR 38.769 is adopted where the CAP duration becomes proportionally shorter with increasing value of M, i.e. if for , duration is  OFDM symbol long, then for , duration is  OFDM symbol long
Note: Duration without CP insertion is considered above, with CP insertion, the total duration may not be exactly proportional
Only following two alternatives for CAP pattern  are considered for further down-selection to one alternative:
Alt 1: ON-OFF-ON-OFF
Alt 2: ON-OFF-ON





R2D Postamble



[Closed] 1st Discussion Round 
FL observations
24 companies provided their views in terms of different options for indicating the end of PRDCH transmission
Overall, 11 companies prefer indicating the end of PRDCH transmission based on some unique pattern that violate Manchester coding rule:
7 companies prefer to explicitly specify postamble design
4 companies prefer to not specify postamble design and leave it up to the reader implementation to violate the Manchester coding rule
8 companies prefer indication the end of PRDCH transmission based on R2D control information. Among the 8 companies there are still different views on whether R2D control information is L1 or not
3 companies consider supporting combination of the above 2 options
From FL perspective, all the options work and doesn’t necessarily have any significant performance difference. However, for R2D control indication based method, it is still not clear whether L1 R2D control information is used or not. Based on the proponents, the views are still diverging. If L1 R2D control information is used, then it needs further discussion on separate or joint CRC attachment. On the other hand, for Manchester coding rule based option, in principle companies have common understanding that violation, for example could be based on transmission of three high or three low voltages, i.e. ON-ON-ON or OFF-OFF-OFF. But the main divergence is whether it needs to be specified or not.


Based on above, proposal 4-1 is provided to take the majority view of indicating end of PRDCH transmission based on violation of Manchester coding rule and also with the consideration that with this option, no further discussion is needed on details related to R2D control information and CRC. For method based on violation of Manchester coding rule, furthermore, it can be discussed whether/what postamble needs to be specified or not. If companies cannot converge whether/what postamble needs to be specified or not, then implementation based solution can be adopted.


(Updated) Proposal 4-1 
Take either Alt 1 or Al 2:
Alt 1: For indicating the end of PRDCH transmission, Manchester coding rule violation corresponding to M value is adopted
FFS: whether the Manchester coding rule violation is based on specified postamble pattern or based on device detecting such violation based on reader’s implementation
Note: If companies cannot converge on whether/what postamble is specified or not, then device detecting such violation based on reader’s implementation will be adopted
Note: If no postamble is specified, then from timing perspective, end of last chip of PRDCH is the end of R2D transmission
Alt 2: There is no consensus to specify R2D postamble



[Closed] 2nd Discussion Round 
[Closed] Proposal 4-1b 
At the end of PRDCH, Manchester coding rule violation corresponding to M value is applied by reader’s implementation that can be used by the device to determine the end of PRDCH transmission
Note: From timing perspective, end of last chip of PRDCH is the end of R2D transmission



Proposal 4-1c 
At the end of PRDCH, Manchester coding rule violation corresponding to M value is applied by reader’s implementation that can be used by the device to determine the end of PRDCH transmission
Note: From timing perspective, end of last chip of PRDCH is the end of R2D transmission



Proposed Conclusion 4-2 
There is no consensus to specify R2D postamble



[Closed] 3rd Discussion Round 
(High Priority) Proposal 4-1d 
Manchester coding rule violation corresponding to M value for PRDCH is applied by reader’s implementation that can be used by the device to determine the end of PRDCH transmission
Note: From timing perspective, end of last chip of PRDCH is the end of R2D transmission
There is no consensus to specify R2D postamble








D2R X-ambles (including preamble & midamble)


[Closed] 1st Discussion Round 
Base Sequence Types: First aspect discussed by companies is related to the preference between the two sequence types including the length that were agreed in RAN1#120bis to be further down-selected and corresponding details including sequence length and generation. In Table 5-1 below, the companies’ views are summarized and also including performance analysis, if provided

Table 5-1: Summary of views between M-sequence and Golay Sequence types for D2R X-ambles



Multiple Preamble lengths: Second aspect discussed by companies is related to the number of preamble types in terms of different lengths. In Table 5-2 below, the companies’ views are summarized and also including performance analysis, if provided

Table 5-2: Summary of views on D2R preamble lengths including long and short preamble



# of Preamble Sequences (with same base sequence): Third aspect discussed by companies is related to the number of sequences of same lengths for preamble/midamble. In Table 5-3 below, the companies’ views are summarized and also including performance analysis, if provided

Table 5-3: Summary of views on number of preamble sequences



1- or 2- part Preamble: Fourth aspect is related to 1-part or 2-part preamble design. In Table 5-4 below, the companies’ views are summarized and also including performance analysis, if provided

Table 5-4: Summary of views on number of parts of preamble



Midamble Configuration: Fifth aspect is related to presence/location of D2R x-ambles including whether pre-defined rule and/or explicit signaling by reader is supported. In Table 5-5 below, the companies’ views are summarized and also including performance analysis, if provided

Table 5-5: Summary of views on midamble configuration





Midamble Sequence: Fifth aspect is related to whether midamble sequence is same as preamble or not. In Table 5-6 below, the companies’ views are summarized and also including performance analysis, if provided

Table 5-6: Summary of views on midamble sequence



FL observations
In terms of the base sequence type, all the companies that provided simulation results showed that the performance is similar with both Golay sequence and M sequence. And in terms of preference, almost there is equal number of companies supporting both options. For M sequence, it was pointed out by few companies that the sequence generation is perhaps simpler compared to Golay sequence. Therefore, FL proposal is to adopt M sequence for D2R x-ambles. 

In terms of preamble types, at least 11 companies prefer to support two preamble formats including short preamble format and long preamble format. Multiple companies pointed out the benefit of supporting the two formats and also, one company demonstrated the benefit of using two lengths. In terms of sequence length, 32(-1) bits is proposed by majority of company. For the shorter format, it was shown that 8(-1) bits have reasonable performance difference, while 16-(1) bits have almost similar performance. Therefore, FL proposes to support both short and long preamble format, with length 8(-1) bits and 32(-1) bits, respectively. Also, all but 2 companies considered same sequence for both midamble and preamble as both serve similar functionality and this would require less specification effort. Therefore, same sequence for both is also proposed by FL. In terms of 1-part or 2-part format for preamble, 4 companies prefer to have 2-part for coarse and fine synchronization, respectively. On the other hand, 2 companies explicitly proposed not to support 2-part preamble. One company showed in evaluations that as long as preamble and midamble are designed optimally in terms of length and sequence, there is not benefit from 2-part preamble. In order to get views from more companies, FL will ask companies to provide inputs on their preference

Regarding the configuration of midambles, ~13 companies prefer explicit indication by the network, while 8 companies prefer pre-defined rule to determine number of midambles and location. In terms of pre-defined rule, majority consider supporting TBS based midamble determination. Furthermore, there are 7 companies prefer a combination of pre-defined rule and explicit indication by network. From FL perspective, all solutions work, but it seems there is a quite good number of companies that prefer additional control to reader in terms of at least the presence and number of midamble due to potentially dynamic change in channel conditions and/or reader-specific implementation. Therefore, FL proposal would be to consider the combined option where the number of midambles can be explicitly indicated by the reader to the device via R2D control information and the position of midambles can be at least pre-defined. 


Based on above summary of views from companies on D2R x-ambles, following proposals are provided




(Updated) Proposal 5-1 
For D2R, 1-part preamble and 1-part midamble, base sequence is generated from M-sequence, where the length of the sequence is 
Value(s) of n = 5, 3
Long preamble is generated based on n = 5
Short preamble is generated based on following n values to be down-selected to single value
Alt 1: n = 3
Alt 2: n = 4
FFS: signaling/application of the two preambles
Same base sequence is applied for both preamble and midamble
Preamble immediately precedes the PDRCH without any gap


[Closed] Question 5-1 
Do you support 1-part or 2-part preamble, where for 2-part preamble, 1st part is based on 0s on 1s without Manchester coding and 2nd part is based on the base sequence (as proposed in proposal 5-1)?


(Updated) Proposal 5-2
For D2R midamble, following two alternatives are considered for further down-selection in RAN1#120bis for the determining the presence and location of midamble(s) at the device:
Alt 1: Reader explicitly indicates the number of midambles to be applied by the device for PDRCH transmission via R2D control information
FFS: details of signaling
Based on the indicated number of midambles, a pre-defined rule is applied at the device to determine the exact location of each of the midamble(s) 
FFS: details of pre-defined rule for determining the exact location of midamble(s) 
Alt 2: Reader explicitly indicates the interval between consecutive x-ambles
FFS: details of signaling
Note: These alternatives don’t preclude support of no midamble




[Closed] 2nd Discussion Round 
[Closed] Proposal 5-1b 
For D2R, 1-part preamble and 1-part midamble, base sequence is generated from M-sequence, where the length of the sequence is 
Value(s) of n
Long preamble is generated based on n = 5
Short preamble is generated based on following n values to be down-selected to single value
Alt 1: n = 3
Alt 2: n = 4
FFS: signaling/application of the two preambles
Same base sequence is applied for both preamble and midamble
Preamble immediately precedes the PDRCH without any gap



Proposal 5-1c 
For D2R preamble/midamble, base sequence is generated from m-sequence, where the length of the sequence is 
Value(s) of n
Long preamble/midamble is generated based on n = 5
Short preamble/midamble is generated based on following n values to be down-selected to single value
Alt 1: n = 3
Alt 2: n = 4
Only 1-part preamble/midamble are supported for D2R
Preamble immediately precedes the PDRCH without any gap
If both long and short preamble/midamble are supported, then reader separately indicates 
For preamble, whether long preamble or short preamble is applied
For midamble, whether long midamble or short midamble is applied
For the same length of preamble and midamble, same sequence is supported


(Updated) Proposal 5-2 
For indicating the interval between consecutive midambles, and between the preamble and the first midamble, via R2D control information, following is adopted:
Unit of interval
Alt 1: Number of “pairs of chips”
Alt 2: Number of bits
FFS: the candidate values in terms of the unit of interval


R-TAS (SIP+CAP)
[Closed] Question 6-1
Based on the agreed SIP pattern and CAP pattern alternatives, could companies indicate their preference for the combined design of SIP + CAP, i.e. R-TAS options below? 
R-TAS Option 1: SIP Alt-1-1, Option 1 + CAP Alt-1
R-TAS Option 2: SIP Alt-1-1, Option 1 + CAP Alt-2
R-TAS Option 3: SIP Alt-1-1, Option 2 + CAP Alt-1
R-TAS Option 4: SIP Alt-1-1, Option 2 + CAP Alt-2
R-TAS Option 5: SIP Alt-2-4, Option 1 + CAP Alt-1
R-TAS Option 6: SIP Alt-2-4, Option 1 + CAP Alt-2
R-TAS Option 7: SIP Alt-2-4, Option 2 + CAP Alt-1
R-TAS Option 8: SIP Alt-2-4, Option 2 + CAP Alt-2



(High Priority) Proposal 6-1v1 
For R-TAS, SIP duration of 1 OFDM symbol is adopted with CAP pattern ON-OFF-ON-OFF for all values of M corresponding to PRDCH 

(High Priority) Proposal 6-1v2 
For R-TAS, SIP duration of 0.5 OFDM symbol is adopted with CAP pattern ON-OFF-ON for all values of M corresponding to PRDCH 



Proposals for offline sessions
1st offline session (Monday, April 7, 2025)
(Updated) Proposal 5-2
For D2R midamble, following two alternatives are considered for further down-selection in RAN1#120bis for the determining the presence and location of midamble(s) at the device:
Alt 1: Reader explicitly indicates the number of midambles to be applied by the device for PDRCH transmission via R2D control information
FFS: details of exact signaling to be discussed under agenda 9.4.4
Based on the indicated number of midambles, pre-defined rule(s) are applied at the device to determine the exact location of each of the midamble(s) 
FFS: details of pre-defined rule for determining the exact location of midamble(s)
Alt 2: Reader explicitly indicates the interval between consecutive x-ambles
FFS: details of exact signaling
Note: These alternatives don’t preclude support of no midamble

(Updated) Proposal 3-1
For CAP of R-TAS, option 1 for CAP of R-TAS from TR 38.769 is adopted with following details:
CAP duration becomes shorter with increasing value of M
FFS: Exact relation between duration of CAP and M values

(Updated) Proposal 2-1 
For the pattern of SIP of R-TAS, only following 2 alternatives are considered for further down-selection to one alternative:
Alt 1-2: ON-OFF with a ratio of 1:3 and with following total SIP duration to be further down-selected during RAN1#120bis:
Option 1: 0.5 OFDM symbol duration
Option 2: 1 OFDM symbol duration
Alt 2-4: ON-OFF-ON-OFF with a ratio of 1:1:1:3 and with following total SIP duration to be further down-selected during RAN1#120bis:
Option 1: 0.5 OFDM symbol duration
Option 2: 1 OFDM symbol duration

Proposal 4-1 
For indicating the end of PRDCH transmission, Manchester coding rule violation corresponding to M value is adopted
FFS: whether the Manchester coding rule violation is based on specified postamble pattern or based on device detecting such violation based on reader’s implementation
Note: If companies cannot converge on whether/what postamble is specified or not, then device detecting such violation based on reader’s implementation will be adopted
Note: If no postamble is specified, then from timing perspective, end of last chip of PRDCH is the end of R2D transmission


(Updated) Proposal 5-1 
For D2R, 1-part preamble and 1-part midamble, base sequence is generated from M-sequence, where the length of the sequence is 
Value(s) of n = 5, 3
Long preamble is generated based on n = 5
Short preamble is generated based on following n values to be down-selected to single value
Alt 1: n = 3
Alt 2: n = 4
FFS: signaling/application of the two preambles
Same base sequence is applied for both preamble and midamble
Preamble immediately precedes the PDRCH without any gap

Proposal 3-2 
For the pattern of CAP of R-TAS, only following 2 alternatives are considered for further down-selection to one alternative:
Alt 1: 3 chips with ON-OFF-ON pattern 
Alt 2: 4 chips with ON-OFF-ON-OFF pattern
Based on the minimum value of M that will be agreed in agenda 9.4.1, corresponding maximum duration of CAP will be selected from the candidate set of {1.5 OFDM symbols, 2 OFDM symbols, 3 OFDM symbols}


2nd offline session (Tuesday, April 8, 2025)
Proposal 5-1b 
For D2R, 1-part preamble and 1-part midamble, base sequence is generated from m-sequence, where the length of the sequence is 
Value(s) of n
Long preamble/midamble is generated based on n = 5
Short preamble/midamble is generated based on following n values to be down-selected to single value
Alt 1: n = 3
Alt 2: n = 4
If both long and short preamble/midamble are supported, reader explicitly indicates it
Same base sequence is applied for both preamble and midamble
Preamble immediately precedes the PDRCH without any gap



Proposal 3-3a
For CAP of R-TAS, following is adopted:
ON-OFF-ON-OFFON pattern with Option 1 for CAP of R-TAS from TR 38.769 is adopted where the CAP duration becomes proportionally shorter with increasing value of M, i.e. if for , duration is  OFDM symbol long, then for , duration is  OFDM symbol long
Note: Duration without CP insertion is considered above, with CP insertion the total duration may not be exactly proportional

Proposal 4-1c 
At the end of PRDCH, Manchester coding rule violation corresponding to M value is applied by reader’s implementation that can be used by the device to determine the end of PRDCH transmission
Note: From timing perspective, end of last chip of PRDCH is the end of R2D transmission


Proposed Conclusion 4-2 
There is no consensus to specify R2D postamble


3rd offline session (Wednesday, April 9, 2025)
Proposal 5-1c 
For D2R preamble/midamble, base sequence is generated from m-sequence, where the length of the sequence is 
Value(s) of n
Long preamble/midamble is generated based on n = 5
Short preamble/midamble is generated based on following n values to be down-selected to single value
Alt 1: n = 3
Alt 2: n = 4
Only 1-part preamble/midamble are supported for D2R
Preamble immediately precedes the PDRCH without any gap
Reader explicitly indicates the following cases to the device
Case of indicating short preamble and short midamble is supported
Case of indicating long preamble and long midamble is supported
[Case of indicating long preamble and short midamble is supported]
[For the same length of preamble and midamble, same sequence is supported]


Proposal 3-3b
For CAP of R-TAS, following is adopted:
ON-OFF-ON-OFFON pattern with Option 1 for CAP of R-TAS from TR 38.769 is adopted where the CAP duration becomes proportionally shorter with increasing value of M, i.e. if for , duration is  OFDM symbol long, then for , duration is  OFDM symbol long
Note: Duration without CP insertion is considered above, with CP insertion, the total duration may not be exactly proportional
Only following two alternatives for CAP pattern  are considered for further down-selection to one alternative:
Alt 1: ON-OFF-ON-OFF
Alt 2: ON-OFF-ON

(Updated) Proposal 5-2 
For indicating the interval between consecutive midambles, and between the preamble and the first midamble, via R2D control information, following is adopted:
Unit of interval
Alt 2: Number of bits after FEC (if FEC is applied) and repetition
FFS: the candidate values in terms of the unit of interval

Proposal 4-1c 
At the end of PRDCH, Manchester coding rule violation corresponding to M value is applied by reader’s implementation that can be used by the device to determine the end of PRDCH transmission
Note: From timing perspective, end of last chip of PRDCH is the end of R2D transmission

Proposed Conclusion 4-2 
There is no consensus to specify R2D postamble




4th offline session (Thursday, April 10, 2025)
(High Priority) Proposal 6-1v1 
For R-TAS, SIP duration of 1 OFDM symbol is adopted with CAP pattern ON-OFF-ON-OFF for all values of M corresponding to PRDCH 
Supporting Companies: Samsung, Vivo, Qualcomm, NTT Docomo, Oppo, IDC, Lenovo, China Telecom, NEC, Tejas, Honor, LGE, CEWit, TCL, Spreadtrum, Futurewei, Ericsson, Panasonic, Fujitsu

(High Priority) Proposal 6-1v2 
For R-TAS, SIP duration of 0.5 OFDM symbol is adopted with CAP pattern ON-OFF-ON for all values of M corresponding to PRDCH 
Supporting Companies: Xiaomi, CMCC

(High Priority) Proposal 4-1d 
Manchester coding rule violation corresponding to M value for PRDCH is applied by reader’s implementation that can be used by the device to determine the end of PRDCH transmission
Note: From timing perspective, end of last chip of PRDCH is the end of R2D transmission
There is no consensus to specify R2D postamble


Proposals for online session
1st online session (Tuesday, April 8, 2025)
Proposal 5-2a
For D2R midamble, following two alternatives are considered for further down-selection in RAN1#120bis for determining the presence and location of midamble(s) at the device:
Alt 1: Reader explicitly indicates the number of midambles to be applied by the device for PDRCH transmission via R2D control information
FFS: details of signaling
Based on the indicated number of midambles, pre-defined rule(s) are applied at the device to determine the exact location of each of the midamble(s) 
FFS: details of pre-defined rule for determining the exact location of midamble(s) 
Alt 2: Reader explicitly indicates the interval between consecutive x-ambles
FFS: details of signaling and any pre-define rule(s), if needed
Note: These alternatives don’t preclude the support of no midamble

Proposal 2-1a
For the pattern of SIP of R-TAS, only following 2 alternatives are considered for further down-selection to one alternative:
Alt 1-2: ON-OFF with a ratio of 1:3 and with following total SIP duration to be further down-selected [during RAN1#120bis]:
Option 1: 0.5 OFDM symbol duration
Note: In this option, the chip duration corresponds to M = 8, but this doesn’t imply support of such reference M value for SIP
Option 2: 1 OFDM symbol duration
Note: In this option, the chip duration corresponds to M = 4, but this doesn’t imply support of such reference M value for SIP
Alt 2-4: ON-OFF-ON-OFF with a ratio of 1:1:1:3 and with following total SIP duration to be further down-selected [during RAN1#120bis]:
Option 1: 0.5 OFDM symbol duration
Note: In this option, the chip duration corresponds to M = 12, but this doesn’t imply support of such reference M value for SIP
Option 2: 1 OFDM symbol duration
Note: In this option, the chip duration corresponds to M = 6, but this doesn’t imply support of such reference M value for SIP

Proposal 3-1a
For CAP of R-TAS, option 1 for CAP of R-TAS from TR 38.769 is adopted with following details:
CAP duration becomes shorter with increasing value of M
FFS: exact relation between duration of CAP and M values including {2,4,6,24}

Proposal 3-2a 
For the pattern of CAP of R-TAS, only following 2 alternatives are considered [for further down-selection to one alternative]:
Alt 1: 3 chips with ON-OFF-ON pattern 
For lowest M  = 2, maximum CAP duration will be 1.5 OFDM symbols with this alternative
Alt 2: 4 chips with ON-OFF-ON-OFF pattern
For lowest M  = 2, maximum CAP duration will be 2 OFDM symbols with this alternative
Based on the minimum value of M that will be agreed in agenda 9.4.2, corresponding maximum duration of CAP will be selected from the candidate set of {1.5 OFDM symbols, 2 OFDM symbols, 3 OFDM symbols}


Proposal 5-1a 
For D2R, 1-part preamble and 1-part midamble, base sequence is generated from M-sequence, where the length of the sequence is 
Value(s) of n = 5, 3
Long preamble is generated based on n = 5
Short preamble is generated based on following n values to be down-selected to single value
Alt 1: n = 3
Alt 2: n = 4
FFS: signaling/application of the two preambles
Same base sequence is applied for both preamble and midamble
Preamble immediately precedes the PDRCH without any gap



Proposal 4-1a 
Take either Alt 1 or Al 2:

Alt 1: For indicating the end of PRDCH transmission, Manchester coding rule violation corresponding to M value is adopted
FFS: whether the Manchester coding rule violation is based on specified postamble pattern or based on device detecting such violation based on reader’s implementation
Note: If companies cannot converge on whether/what postamble is specified or not, then device detecting such violation based on reader’s implementation will be adopted
Note: If no postamble is specified, then from timing perspective, end of last chip of PRDCH is the end of R2D transmission
Alt 2: There is no consensus to specify R2D postamble

2nd online session (Wednesday, April 9, 2025)
Proposal 5-1d 
For D2R preamble/midamble, base sequence is generated from m-sequence, where the length of the sequence is 
Value(s) of n
Long preamble/midamble is generated based on n = 5
Short preamble/midamble is generated based on following n values to be down-selected to single value
Alt 1: n = 3
Alt 2: n = 4
Only 1-part preamble/midamble are supported for D2R
Preamble immediately precedes the PDRCH without any gap
If both long and short preamble and midamble (if present) are supported, then following cases are supported and reader explicitly  indicates one of the following cases for PDRCH:
Short preamble and short midamble 
Long preamble and long midamble 
[Long preamble and short midamble]
[For the same length of preamble and midamble, same sequence is supported]

Proposal 3-3c
For CAP of R-TAS, following is adopted:
Option 1 for CAP of R-TAS from TR 38.769 is adopted where the CAP duration becomes proportionally shorter with increasing value of M, i.e. if for , duration is  OFDM symbol long, then for , duration is  OFDM symbol long
Note: Duration without CP insertion is considered above, with CP insertion, the total duration may not be exactly proportional
Only following two alternatives for CAP pattern  are considered for further down-selection to one alternative:
Alt 1: ON-OFF-ON-OFF
Alt 2: ON-OFF-ON

Proposal 5-2a 
For indicating the interval between consecutive midambles, and between the preamble and the first midamble, via R2D control information, following is adopted:
Unit of interval is number of bits after FEC (if FEC is applied) and repetition
FFS: the candidate values in terms of the unit of interval
Proposal 4-1c 
At the end of PRDCH, Manchester coding rule violation corresponding to M value is applied by reader’s implementation that can be used by the device to determine the end of PRDCH transmission
Note: From timing perspective, end of last chip of PRDCH is the end of R2D transmission

Proposed Conclusion 4-2 
There is no consensus to specify R2D postamble


3rd online session (Thursday, April 10, 2025)
Either Take v1 or v2 for Proposal 6-1
(High Priority) Proposal 6-1v1 
For R-TAS, SIP duration of 1 OFDM symbol is adopted with CAP pattern ON-OFF-ON-OFF for all values of M corresponding to PRDCH 
Supporting Companies (19 companies): Samsung, Vivo, Qualcomm, NTT Docomo, Oppo, IDC, Lenovo, China Telecom, NEC, Tejas, Honor, LGE, CEWit, TCL, Spreadtrum, Futurewei, Ericsson, Panasonic, Fujitsu

(High Priority) Proposal 6-1v2 
For R-TAS, SIP duration of 0.5 OFDM symbol is adopted with CAP pattern ON-OFF-ON for all values of M corresponding to PRDCH 
Supporting Companies (2 companies): Xiaomi, CMCC

(High Priority) Proposal 4-1d 
Manchester coding rule violation corresponding to M value for PRDCH is applied by reader’s implementation that can be used by the device to determine the end of PRDCH transmission
Note: From timing perspective, end of last chip of PRDCH is the end of R2D transmission
There is no consensus to specify R2D postamble








Contributions in RAN1#120bis


Appendix
Revised WID (RP-243326): RAN1 Scope & Objectives 
General Scope
The definitions provided in TR 38.848, TR 38.769, and decisions, etc. made during the Rel-19 SI in RAN WGs are taken into this WI, and the following is the exclusive general scope:
The overall objective shall be to standardize the following Ambient IoT device:
Device 1: ~1 µW peak power consumption, has energy storage, RF envelope detector receiver, initial sampling frequency offset (SFO) up to 10X ppm, neither R2D nor D2R amplification in the device. The device’s D2R transmission is backscattered on a carrier wave provided externally.
Deployment scenario 1 with Topology 1, according to D1T1-B. 
FR1 licensed spectrum in FDD, with R2D in DL spectrum and D2R and CW in UL spectrum.
Spectrum deployment in-band to NR and standalone, with A-IoT BS located indoor.
Traffic types DO-DTT, DT, for rUC1 (indoor inventory) and rUC4 (indoor command). 
Carrier wave transmission for waveform 1 only, without hopping, per the following cases in TR 38.769:
Case 1-4 for D1T1-B
Proximity determination via Solution 1 in TR 38.769 only.
Device (un)availability via Direction 1 in TR 38.769 only.

WGs begin their discussions from the decisions already made in TR 38.769, with the following refinements for the scope: 

The following objectives are set, within the General Scope:
RAN1 scope:
PRDCH and PDRCH, which are the only physical channels in R2D and D2R, respectively.
R2D and D2R signal(s)
Multiplexing/multiple access in R2D is by only TDMA, and in D2R is by only TDMA and FDMA.
R2D supports only OOK-4 modulation, one solution for CP handling. D2R backscattering supports only OOK and BPSK modulations.
R2D transmission supports only the Manchester line code in TR 38.769
D2R transmission supports:
Either the Manchester line code in TR 38.769 or no line code (one to be down-selected); and
A corresponding small frequency shift method according to the options in TR 38.769.
R2D does not support FEC. D2R supports only convolutional code with generator polynomials as per TS 36.212. Applying or not applying the FEC to D2R is specified by ensuring it is under the reader control and applies to all devices targeted by the reader.
PRDCH and PDRCH both support transmission without CRC, and with CRC as per the generator polynomials in TS 38.212 for 6-bit CRC and 16-bit CRC. Cases to use which length of CRC, or no CRC, to be decided in RAN1.
D2R supports physical layer repetition transmission. R2D does not support physical-layer repetition transmission. 
RAN2 scope:
Specify the necessary functions and procedures for an Ambient IoT compact protocol stack and lightweight signalling procedure to enable DO-DTT and DT data transmission:
A-IoT Paging, including subsequent paging for the same service. Support the options that a paging message contains one identifier, and that a paging message contains no identifier. 
Note: RAN2 aims to design a paging message format such that multiple identifiers can be contained in one paging message, for forward compatibility purposes.
A-IoT Random access, including re-access for failure handling. Contention-based and contention-free cases are supported. For the contention-based random access, only Solution 1 (3-step only) is included.
A-IoT data transmission, including data (re-)transmission for failure handling. Segmentation is supported at least in D2R.
Only MAC layer is included
RAN3 scope: 
Specify necessary architectural aspects, and signaling and procedures between A-IoT RAN and A-IoT CN to support the A-IoT functions, assuming an architecture of aggregated gNB, including:
Inventory and command operations
Device location reporting at reader ID granularity
Note: The above A-IoT functions are supported over the existing NG interface, based on architecture(s) defined by RAN3/SA2.
RAN4 scope:
Specify RF requirements for Ambient-IoT BS, device 1, and CW
RF requirements for Type 1-C Ambient-IoT BS
RF requirements for device 1
RF requirements for CW
Specify RRM core requirements for device 1, if necessary
Study and develop OTA test methodology for A-IoT device 1
Consider test methods specified in TR 38.870 as starting point. Take test system reuse, test system complexity and test time into account, when developing test methods suitable for Ambient IoT.
Develop the preliminary Measurement Uncertainty (MU) assessment for the test system
Use band n8 as an example band

Note 1: Coordination with SA2 and SA3 is expected. Updates to the WID objectives should be considered if needed.

Note 2: This WI shall target for an IoT segment well below the existing 3GPP IoT technologies, e.g. NB-IoT, eMTC, RedCap, etc. The WI shall not aim to replace existing 3GPP LPWA technologies.

SI Phase: RAN1 Agreements (relevant for R2D/D2R signals including timing acquisition and synchronization
RAN1#116 (Athens, Greece, February 26th – March 1st, 2024)
Agreement
At least the following time domain frame structure is studied for A-IoT R2D and D2R transmission.
For R2D transmission,
A R2D timing acquisition signal (e.g. R2D preamble) is included at least for timing acquisition and for indicating the start of the R2D transmission in time domain.
For D2R transmission,
A D2R timing acquisition signal (e.g. D2R preamble) is included at least for timing acquisition and for indicating the start of the D2R transmission in time domain.
FFS other necessary component(s), e.g. midamble, postamble, periodic sync signal, control fields, guard period


RAN1#116bis (Changsha, Hunan Province, China, April 15th – April 19th, 2024)

Agreement
To determine or derive the end of PRDCH transmission, study at least following options:  
Option 1: R2D postamble immediately follows the PRDCH to indicate the end of the PRDCH.       
Option 2: Based on R2D control information.

Agreement
For the reader to acquire the end of PDRCH transmission, study at least following options:  
Option 1: D2R postamble immediately follows the PDRCH
Option 2: Based on control information

Agreement
For D2R transmission, study the necessity of midamble at least for the purpose of performing timing/frequency tracking or channel estimation or interference estimation, considering at least the following: 
Modulation and Coding schemes, e.g., data modulation, line/channel coding 
Receiving methods, e.g., coherent or non-coherent
D2R transmission length/packet size
Midamble overhead
Timing/frequency accuracy
Phase accuracy

Agreement
RAN1 study the R2D transmission without midamble as the baseline if Manchester encoding is used.
FFS the necessity for the R2D transmission with midamble if PIE is used. 

Agreement
For the R2D timing acquisition signal immediately preceding the transmission of a physical channel, study a preamble with at least two parts which includes a start-indicator part and a clock-acquisition part, where the start-indicator part immediately precedes the clock-acquisition part:
Start-indicator part provides the start of the R2D transmission
FFS: Details of start-indicator part
Clock-acquisition part provides at least the chip synchronization of the subsequent physical channel transmission
FFS: Details of clock-acquisition part, e.g. structure, encoding, length, etc. 
FFS: Methods to determine chip duration of the subsequent physical channel transmission 
FFS: Other functionalities
Note: the preamble is considered not to be part of a physical channel
FFS: other part(s) of the preamble, if any 
FFS: whether the above clock acquisition is sufficient for all devices
FFS: how to make the preamble compact

Agreement
For D2R, a preamble preceding each PDRCH transmission is studied as the baseline at least for the D2R timing acquisition signal:
Preamble is not part of PDRCH
FFS: Other functionalities of the preamble

Agreement
Reference signals including at least DMRS, PTRS, CSI-RS/TRS, are not further studied for R2D.

Agreement
Reference signals including DMRS, PTRS, SRS, are not further studied for D2R
Note: This doesn’t preclude the possibility to study preamble, midamble, postamble for different purposes, e.g. channel/interference estimation and/or proximity determination


RAN1#117 (Fukuoka City, Fukuoka, Japan, May 20th – 24th, 2024)

Agreement
For the start-indicator part of the R2D time acquisition signal, study the two options below:
Option 1: ON/OFF pattern i.e. high/low voltage transmission 
Option 2: OFF pattern, i.e. low voltage transmission 

Agreement
For R2D, the clock-acquisition part of the R2D time acquisition signal is used to determine the OOK chip duration
FFS: Pattern design to support determination of chip duration


RAN1#118 (Maastricht, NL, August 19th – 23rd,  2024)

Agreement
For each D2R transmission, no separate part for start-indicator is considered for the preamble preceding the PDRCH.

Agreement
For D2R transmission, preamble preceding the PDRCH is studied also for the potential additional functionalities:
SFO estimation
CFO estimation
Channel estimation
Interference estimation
Note: this does not preclude studying the above functionalities by using a midamble and/or postamble, if supported
FFS: Other functionalities, if any

Agreement
For the start-indicator part of the R2D time acquisition signal, ON/OFF pattern i.e. high/low voltage transmission is applied
FFS: length/pattern of ON/OFF.
FFS: when TD2R_min is applicable, whether/how the start-indicator part is included in TD2R_min or not. To be discussed in 9.4.2.2


RAN1#118bis (Hefei, China, October 14th – 18th,  2024)

Agreement
The start indicator part of the R2D time acquisition signal is not included in TD2R_min.

Agreement
The TR will capture the following options, and companies are encouraged to analyze the tradeoffs among the following D2R amble(s) options:
Option 1: D2R preamble only
Option 2: D2R preamble + X midamble(s), where X 1
Option 3: D2R preamble + postamble
Option 4: D2R preamble + Y midamble(s) + postamble, where Y1
For the above options, companies are encouraged to report at least the following:
Purpose(s) of the preamble, midamble and postamble 
Whether companies assume multiple options can be supported


Agreement
For analysing the trade-offs among the D2R amble(s) options, companies can refer to the Table 3.2.4 in section 3.2.4 of R1-2408993 for information.

Agreement
For the clock-acquisition part of the R2D time acquisition signal, following is captured in the TR 38.769:
Clock-acquisition part is based on OOK without line coding and includes rising/falling edges, including at least two rising or at least two falling edges for the device to determine the OOK chip duration

Agreement
For the start-indicator part of the R2D time acquisition signal, for providing the start of the R2D transmission, following is captured in the TR 38.769:
Following options have been studied for the start-indicator part of the R2D time acquisition signal:
Option 1: ON-OFF transmission is considered based on energy/edge detection, and multiple alternatives have been studied including 
Alt 1: A single ON-OFF transmission, i.e. one high-voltage transmission followed by one low-voltage transmission, where ON and OFF may have same or different durations
Alt 2: A multi-ON-OFF transmission, where different ON and different OFF may have same or different durations and different parts may have same or different duration
Option 2: ON-OFF sequence-based design is considered which consists of a pre-defined sequence for detection of start-indicator part based on digital correlation
For both the options, it is observed that a fixed duration for the start-indicator part can be considered, regardless of the value of M used for PRDCH transmissions. 
Miss-detection ratio (MDR), false-alarm ratio (FAR) and detection complexity have been considered for the design of the R2D start indicator part by following companies
It is observed by 1 source [Huawei] that for an FAR of ~0%, the MDR of less than 1% can be achieved with Alt 2 of option 1 (considering 2 ON-OFF transmissions with different durations) and it is also observed that low-complexity and reduced power consumption can be achieved
1 source [ZTE] evaluated Alt 1 of option 1 (considering same duration for ON and OFF) and Alt 2 of option 1 (considering multiple ON-OFF transmissions with same duration) and observed that for an FAR of ~0%, the MDR of less than 1% can be achieved and Alt 1 of option 1 performs better than Alt 2 of option 1. 
1 source [CATT] observed with ON-OFF pattern, that for an FAR of ~0%, the MDR of less than 1% can be achieved with a duration of at least 1 OFDM symbol
1 source [Qualcomm] compares the performance between option 1 and option 2. It shows almost similar coverage range (SNR requirement) for target MDR of 1%. For MDR of 10%, it shows that sequence-based design provides better performance, and it is observed that during the available time, it is feasible for all devices to detect the start-indicator sequence within the power budget. It is further observed that the FAR with sequence-based design can be improved in case of interference scenarios when compared with pattern-based design. 
For both the options, it may be beneficial that the start-indicator part is distinguishable at least from other parts of the R2D transmissions

Agreement
For the clock-acquisition part of the R2D time acquisition signal for OOK chip duration determination, following options are studied:
Option 1: Duration of the clock-acquisition part is variable for different M values, i.e. the duration becomes shorter with increasing value of M
Option 2: Duration of the clock-acquisition part is constant for different M values based on repetition, i.e. repetition factor is increased with increasing value of M to keep the duration constant
FFS: Whether/what restriction on M values for the clock-acquisition part
Note: Other functionalities of clock-acquisition part is a separate discussion

Agreement
For the D2R preamble, binary signal is considered.



RAN1#119 (Orlando, US, Nov 18th – 22nd, 2024)
Agreement
Capture following observations in the TR 38.769, where CFO is assumed to be zero or negligible.
For coherent detection of PDRCH with a payload of 16 bits or 20 bits with 6-bit or 16-bit CRC, using 1/2 Manchester coding and 1/3 or 1/2 convolutional code:
Source [6, ZTE] observed that with up to 10% SFO, ~1kbps data rate, when the same amble(s) overhead is maintained, Option 3 provides comparable performance results to Option 1.
Source [7, Samsung] observed that with up to 10% SFO, ~5kbps data rate, for device 1 and with up to 1% SFO for device 2, the decoding performance with/without midamble are similar
Source [9, vivo] observed that Option 1 is sufficient to achieve 10% and 1% BLER, with no more than 8 SFO hypotheses tested at the reader side.
With up to 10% SFO, ~ 5kbps data rate, the SNR needed to achieve 10% and 1% BLER is similar (~ -2dB and 4 dB) for Option 1, Option 2 of D2R preamble+1midamble and Option 3.
With up to 1% SFO, ~ 5kbps data rate, the SNR needed to achieve 10% and 1% BLER is similar (~ -2.8dB and 3.3dB) for Option 1, Option 2 of D2R preamble+1 midamble and Option 3.
Source [20, OPPO] observed that with perfect SFO estimation, 1kbps data rate and OOK modulation, there is no noticeable performance gain from using midamble(s) and/or postamble for PDRCH transmission, regardless of the payload size.
Source [30, Qualcomm] observed that the required SFO estimation accuracy to achieve 1% and 10% BLER is <5*10^3 ppm
With up to 10% SFO, achieving the required accuracy necessitates more than 20 SFO hypotheses at the reader side for Option 1 and 10 SFO hypotheses are sufficient for Option 3 of D2R preamble + postamble. But for Option 3 reader has to store the received samples and wait for the postamble that is after the end of PDRCH for any of SFO/channel/interference estimation, demodulation, decoding, etc. 
With up to 1% SFO, 4 SFO hypotheses are sufficient for Option 1 to achieve the required accuracy.

For coherent detection of PDRCH with a payload of 96bits with 16-bit CRC (or 6-bit CRC [14, Xiaomi]), using 1/2 Manchester coding and 1/3 or 1/2 convolutional code,
Sources [3, Huawei], [5, CMCC] and [14, xiaomi] observed that Option 1 cannot achieve 10% BLER.
Sources [6, ZTE], [7, Samsung], [9, vivo], [20, OPPO] and [30, QC] observed that Option 1 can achieve 10% BLER.
Sources [3, Huawei], [5, CMCC], [6, ZTE], [7, Samsung], [9, vivo], [14, xiaomi], [16, China Telecom] observed that adding additional amble improves the performance. 
Source [3, Huawei] observed that with up to 10% SFO, 
Option 2 of D2R preamble+ 1 midamble achieves 10% BLER at SNR around -3dB, but cannot achieve 1% BLER.
Option 3 of D2R preamble+ postamble achieves 10% BLER at SNR around -4dB, and can achieve 1% BLER at SNR around 4dB.
Source [5, CMCC] observed that with up to 10% SFO, Option 3 allows reader to precisely search and detect the SFO with 0.03% residual SFO at -3dB SNR TDL-A channel, achieving 10% BLER -2.44dB SNR for ~1 kbps data rate and -2.17 dB for ~6 kbps data rate. Source [5, CMCC] further observed that when the reader adopts same number of SFO hypothesis based on preamble, with 1% SFO, Option 3 can achieve 10% BLER at -4.27 dB SNR for ~1 kbps and at -4.29 dB SNR for ~6 kbps, which provides 1~2 dB performance gain when compared to 10% SFO. 
Source [14, xiaomi] observed that with up to 10% SFO, Option 2 of D2R preamble + 1 midamble, option 3, and option 4 of D2R preamble + 1 midamble+postamble achieve basically the same performance, the SNR for 10% BLER is 5dB for 1.25 kbps data rate.  
Source [6, ZTE] observed that with up to 10% SFO, ~1kbps data rate, and the same amble(s) overhead, Option 3 can provide 1~2 dB, 5dB performance gain for 10% BLER, 1% BLER, respectively, compared to Option 1. Additionally, Option 3 can provide ~1dB, 2dB performance gain for 10% BLER, 1% BLER, respectively, compared to Option 2 of D2R preamble +1 midamble.
Source [7, Samsung] observed that for ~5kbps data rate, compared to option 1, 
For device 1 with up to 10% SFO, Option 2 of D2R preamble + 1 midamble provides ~0.5 dB SNR gain at 10% BLER with TDL-A channel and ~0.9 dB SNR gain with TDL-D channel.
For device 2 with up to 1% SFO, Option 2 of D2R pramble + 1midamble provides ~1 dB SNR gain at 10% BLER with TDL-A channel and ~1.4 dB SNR gain with TDL-D channel.
Source [9, vivo] observed that, 
With up to 10% SFO, ~5.5kbps data rate, Option 1 achieves 10% and 1% BLER at SNR ~0.7dB and 10dB, respectively; Additionally, maintaining the same amble overhead, Option 2 (D2R preamble + 1 midamble) and Option 3 demonstrate similar performance, achieving 10% and 1% BLER at SNR around -1.7dB and 5.2dB, respectively.
With up to 1% SFO, ~5.5kbps data rate, Option 1 achieves 10% and 1% BLER at SNR ~ -1.3dB and 11dB, respectively. Additionally, with the same amble overhead, the SNR difference between Option 2 (D2R preamble+1midamble) and Option 3 for 10% and 1% BLER is less than 1dB, with SNRs ~ -3.1dB to -2.5dB for 10% BLER and ~3.6dB to 4.5dB for 1% BLER.
Source [16, China Telecom] observed that with up to 10% SFO, ~7.5kbps data rate, there is ~6~7dB performance gap at 10% BLER and ~10.5~11.5dB performance gap at 1% BLER between option 2 of D2R preamble+111 midambles and option 1. Note that Source [16, China Telecom] does not use any convolutional code.
Source [20, OPPO] observed that with perfect SFO estimation, 1kbps data rate and OOK modulation, there is no noticeable performance gain from using midamble(s) and/or postamble for PDRCH transmission, regardless of the payload size.
Source [30, Qualcomm] observed that the required SFO estimation accuracy to achieve 1% and 10% BLER is <10^3 ppm. To achieve the required accuracy,
For Option 1, more than 50 SFO hypotheses at reader side are necessary for device with up to 10% SFO and 6 SFO hypotheses are sufficient at reader side for device with up to 1% SFO. 
For Option 3, 10 SFO hypotheses are sufficient for device with up to 10% SFO, but reader has to store the received samples and wait for the postamble that is after the end of PDRCH for any of SFO/channel/interference estimation, demodulation, decoding, etc.

For coherent detection of PDRCH with a payload of 400bits with 16-bit CRC, using 1/2 Manchester coding and 1/3 or 1/2 convolutional code, 
For option 1 of D2R preamble only, 
Sources [3, Huawei], [5, CMCC], [6, ZTE], [8, Spreadtrum], [9, vivo], [14, xiaomi] observed that with up to 10% SFO, 10% BLER cannot be achieved. 
Source [20, OPPO] observed that with perfect SFO estimation, 1kbps data rate and OOK modulation, there is no noticeable performance gain from using midamble(s) and/or postamble for PDRCH transmission.
For other amble options, 
Source [3, Huawei] observed that
With accurate SFO estimation, Option 2 of D2R preamble + 4 midambles can achieve 10% BLER at SNR ~ 2.7dB but cannot achieve 1% BLER.
With up to 10% SFO, Option 3 cannot achieve 10% BLER.
With up to 10% SFO, Option 4 of D2R preamble+2 midambles+postamble achieves 10% BLER at SNR of ~0.25dB; But it cannot achieve 1% BLER. Option 4 of D2R preamble+3 or 4 midambles+postamble, achieves a 10% BLER at an SNR of around -0.2 dB, and achieves 1% BLER at SNR around 9dB or 8dB, respectively.
Source [5, CMCC] observed that with up to 10% SFO, Option 4 of D2R preamble combined with 1 to 4 midambles + postamble, achieves 10% BLER at SNR of 2.5 dB, 1 dB, 0.8 dB, or 0.5 dB, respectively, for a data rate of around 1 kbps.
Source [6, ZTE] observed that with up to 10% SFO, 
Option 3 can provide ~5.5 dB performance gain compared to option 2 of D2R preamble+1midamble for 10% BLER, with the same amble(s) overhead for ~1kbps data rate.
Option 2 of D2R preamble+1midamble cannot achieve 1% BLER for ~1kbps data rate.
Option 4 of the D2R preamble+1 or 2 midamble(s)+postamble, has similar performance, it can achieves a 10% BLER at SNR of -1dB and achieves a 1% BLER at SNR of 6dB and 5dB respectively for ~1kbps data rate.  
Source [8, Spreadtrum] observed that with up to 10% SFO, 
Option 3 of D2R preamble+ postamble cannot achieve 10% BLER for ~7kpbs.  
Option 4 of D2R preamble + 1 midamble + postamble can achieve 10% BLER and 1% BLER at SNR around -6dB and 0 dB, respectively for ~7kpbs data rate.
Source [9, vivo] observed that 
With up to 10% SFO, maintaining the same amble overhead, both Option 2 of D2R preamble+5 midambles and Option 4 of D2R preamble+4 midambles+1 postamble demonstrate similar performance, achieving 10% BLER at SNR ~0.2dB. While for 1% BLER, the SNR for Option 2 and Option 4 is ~ 9.2dB and 12.8dB, respectively for ~5.5kpbs data rate.   
With up to 1% SFO, maintaining the same amble overhead, both Option 2 of D2R preamble+5 midambles and Option 4 of D2R preamble+4 midambles+1postamble demonstrate similar performance, achieving 10% BLER at SNR around -1.2dB. While for 1% BLER, the SNR for Option 2 and Option 4 is ~7.8dB and 9.1dB, respectively for ~5.5kpbs data rate.   
Source [14, xiaomi] observed that with up to 10% SFO, Option 2 of D2R preamble+3 midambles and Option 4 of D2R preamble+3 midambles+postamble can achieve 10% BLER when the SNR is within the range of [15, 25] dB for 1.25 kbps data rate. 
Source [30, Qualcomm] observed that the required SFO estimation accuracy to achieve 1% and 10% BLER is much smaller than 10^3 ppm. To achieve the required accuracy,
For Option 2 of D2R preamble+X midamble(s) where midamble inserted per every certain number of PDRCH bits (e.g., 192 bits),
For SFO estimation using each amble for the subsequent PDRCH bits (e.g., 192 bits), with up to 10% SFO, more than 50 SFO hypotheses are necessary at the reader side and with up to 1% SFO, 6 SFO hypotheses are sufficient at the reader side.
For SFO estimation based on the time gap between preamble and midamble, with up to 10% SFO, 10 SFO hypotheses are used, but reader has to store the received samples and wait for the midamble to start SFO/channel/interference estimation, demodulation, decoding, etc.
For Option 3 of D2R preamble+postamble, SFO estimation is based on the time gap between preamble and postamble, with up to 10% device SFO, 10 SFO hypotheses are used for reader, but reader has to store the received samples and wait for the postamble that is after the end of PDRCH for any of SFO/channel/interference estimation, demodulation, decoding, etc

For the synchronization and timing tracking of D2R transmission, 
Source [5, CMCC] report that with up to 10% SFO, option 1 is not sufficient for D2R reception since the residual SFO at reader side is larger than 1%. While with option 3, the reader can precisely search and detect the SFO with a residual SFO of 0.03% at -3dB SNR TDL-A channel.
Source [14, xiaomi] report that 
For packet size of 96bits, when the SNR is increased from -4dB to 20dB, the ratio of device residual SFO over 100ppm decreases to 6% for Option 2, 3 and 4, but remains at 95% for Option 1.
For packet size of 400bits, when the SNR is increased from -4dB to 20dB, the ratio of device residual SFO larger than 10ppm decreases to 5% for Option 2, 3, and 4, but is still 99.6% for Option 1.
Sources [9, vivo], [15, CATT] report that SFO estimation based on D2R preamble can achieve accurate estimation without additional ambles (midamble or postamble). 
Source [9, vivo][7 Samsung] observed that for non-coherent detection of PDRCH, the number of SFO hypotheses and the SNR needed for 10% and 1% BLER cannot significantly be reduced for option 2, 3 and 4 compared to the option 1. Moreover, the additional ambles i.e., midamble(s) and/or postamble introduces additional overhead and prevents pipelined processing of the reception.     
Source [15, CATT] observed that 
The coarse estimation of SFO based on the D2R preamble indicates that the SFO estimation error is less than 1% with a probability of 99.3%, and less than 0.1% with a probability of 49.9%.
The fine estimation of SFO based on the D2R preamble shows that the SFO estimation error is less than 1% with a probability of 99.5%, and less than 0.1% with a probability of 90.8%.
Reader/gNB can achieve a probability of not less than 99.5% for SFO estimation error below 1%, and 90.8% for SFO estimation error below 0.1% by receiving D2R preamble signals.
Source [30, Qualcomm] report that for D2R with coherent demodulation at reader, the reader needs to estimate the device clock frequency with the accuracy of 0.5% (5 * 10^3 ppm) or lower for a short message (e.g., 72 bits after CRC/coding) and 0.1% (10^3 ppm) or lower for a long message (e.g., 224 bits after CRC/coding). The source further reports that design of D2R amble(s) (e.g., overhead) and the correspondingly required number of SFO hypothesis for the estimation depend on the sampling clock accuracy that the device uses for D2R. 
Note: in the observations above where coherent detection is used, sources that evaluated option 3 and option 4 assumed that the postamble is used at least for time/frequency tracking and for channel estimation.


Agreement
For the CFO calibration signal, which is required only for device 2b to reduce the frequency offset range and the guard-bandwidth of D2R transmission, the following observations are captured in TR 38.769:
Source [3, Huawei] report that a single-tone RF signal is used as the CFO calibration signal, it is not a part of time acquisition signal and can be transmitted as an optional R2D signal after the PRDCH transmission. 
Sources [2, Ericsson], [19, Panasonic] and [20, OPPO] report that additional synchronization signal is needed. 
[OPPO] state the R2D timing acquisition signal may not be sufficient or may not be usable for CFO calibration since a reference frequency is needed when separate LOs are used for Tx and Rx in device 2b.
Sources [7, Samsung], [9, vivo], [30, Qualcomm], [36, Apple] report that additional synchronization signal is needed if the synchronization for carrier frequency using R2D signal/channel does not provide required functionalities for device 2b.
Source [5, CMCC][31, MTK] report that it may not be possible to achieve enough frequency accuracy (0.01 ppm) even after CFO calibration based on R2D time acquisition signals for coherent detection at reader especially when the D2R data rate is low.

Agreement
For device 2b, a signal for CFO calibration should be provided to synchronize / calibrate the device clock for LO for carrier frequency (Clock purpose #5) to achieve the accuracy after clock sync / calibration at device side captured in Table 5.2.3-1.
Frequency calibration at device 2b is beneficial at least to reduce the guard-bandwidth of D2R transmission.

Agreement
Adopt the updates documented in R1-2410653 for section 6.2 of the TR38.769. 

Agreement
Adopt following update to the TP agreed on Monday

Capture following observations in the TR 38.769, where CFO is assumed to be zero or negligible.
[omit unchanged part]
For coherent detection of PDRCH with a payload of 96bits with 16-bit CRC (or 6-bit CRC [14, Xiaomi]), using 1/2 Manchester coding and 1/3 or 1/2 convolutional code,
Sources [3, Huawei], [5, CMCC] and [14, xiaomi] observed that Option 1 cannot achieve 10% BLER.
Sources [6, ZTE], [7, Samsung], [9, vivo], [20, OPPO] and [30, QC] observed that Option 1 can achieve 10% BLER.
Sources [3, Huawei], [5, CMCC], [6, ZTE], [7, Samsung], [9, vivo], [14, xiaomi], [16, China Telecom] observed that adding additional amble improves the performance. 
Source [3, Huawei] observed that with up to 10% SFO, 
Option 2 of D2R preamble+ 1 midamble achieves 10% BLER at SNR around -3dB, but cannot achieve 1% BLER.
Option 3 of D2R preamble+ postamble achieves 10% BLER at SNR around -4dB, and can achieve 1% BLER at SNR around 4dB.
Source [5, CMCC] observed that with up to 10% SFO, Option 3 allows reader to precisely search and detect the SFO with 0.03% residual SFO at -3dB SNR TDL-A channel, achieving 10% BLER -2.44dB SNR for ~1 kbps data rate and -2.17 dB for ~6 kbps data rate. Source [5, CMCC] further observed that when the reader adopts same number of SFO hypothesis based on preamble, with 1% SFO, Option 3 can achieve 10% BLER at -4.27 dB SNR for ~1 kbps and at -4.29 dB SNR for ~6 kbps, which provides 1~2 dB performance gain when compared to 10% SFO. 
Source [14, xiaomi] observed that with up to 10% SFO, Option 2 of D2R preamble + 1 midamble, option 3, and option 4 of D2R preamble + 1 midamble+postamble achieve basically the same performance, the SNR for 10% BLER is 5dB for 1.25 kbps data rate.  
Source [6, ZTE] observed that with up to 10% SFO, ~1kbps data rate, and the same amble(s) overhead, Option 3 can provide 1~2 dB, 5dB performance gain for 10% BLER, 1% BLER, respectively, compared to Option 1. Additionally, Option 3 can provide ~1dB, 2dB performance gain for 10% BLER, 1% BLER, respectively, compared to Option 2 of D2R preamble +1 midamble.
Source [7, Samsung] observed that for ~5kbps data rate, compared to option 1, 
For device 1 with up to 10% SFO, Option 2 of D2R preamble + 1 midamble provides ~0.5 dB SNR gain at 10% BLER with TDL-A channel and ~0.9 dB SNR gain with TDL-D channel.
For device 2 with up to 1% SFO, Option 2 of D2R pramble + 1midamble provides ~1 dB SNR gain at 10% BLER with TDL-A channel and ~1.4 dB SNR gain with TDL-D channel.
Moreover, the additional ambles i.e., midamble(s) and/or postamble introduces additional overhead and postamble may prevent pipelined processing of the reception. 
Source [9, vivo] observed that, 
With up to 10% SFO, ~5.5kbps data rate, Option 1 achieves 10% and 1% BLER at SNR ~0.7dB and 10dB, respectively; Additionally, maintaining the same amble overhead, Option 2 (D2R preamble + 1 midamble) and Option 3 demonstrate similar performance, achieving 10% and 1% BLER at SNR around -1.7dB and 5.2dB, respectively.
With up to 1% SFO, ~5.5kbps data rate, Option 1 achieves 10% and 1% BLER at SNR ~ -1.3dB and 11dB, respectively. Additionally, with the same amble overhead, the SNR difference between Option 2 (D2R preamble+1midamble) and Option 3 for 10% and 1% BLER is less than 1dB, with SNRs ~ -3.1dB to -2.5dB for 10% BLER and ~3.6dB to 4.5dB for 1% BLER.
Source [16, China Telecom] observed that with up to 10% SFO, ~7.5kbps data rate, there is ~6~7dB performance gap at 10% BLER and ~10.5~11.5dB performance gap at 1% BLER between option 2 of D2R preamble+111 midambles and option 1. Note that Source [16, China Telecom] does not use any convolutional code.

Source [20, OPPO] observed that with perfect SFO estimation, 1kbps data rate and OOK modulation, there is no noticeable performance gain from using midamble(s) and/or postamble for PDRCH transmission, regardless of the payload size.
Source [30, Qualcomm] observed that the required SFO estimation accuracy to achieve 1% and 10% BLER is <10^3 ppm. To achieve the required accuracy,
For Option 1, more than 50 SFO hypotheses at reader side are necessary for device with up to 10% SFO and 6 SFO hypotheses are sufficient at reader side for device with up to 1% SFO. 
For Option 3, 10 SFO hypotheses are sufficient for device with up to 10% SFO, but reader has to store the received samples and wait for the postamble that is after the end of PDRCH for any of SFO/channel/interference estimation, demodulation, decoding, etc.

For coherent detection of PDRCH with a payload of 400bits with 16-bit CRC, using 1/2 Manchester coding and 1/3 or 1/2 convolutional code, 
For option 1 of D2R preamble only, 
Sources [3, Huawei], [5, CMCC], [6, ZTE], [8, Spreadtrum], [9, vivo], [14, xiaomi] observed that with up to 10% SFO, 10% BLER cannot be achieved. 
Source [20, OPPO] observed that with perfect SFO estimation, 1kbps data rate and OOK modulation, there is no noticeable performance gain from using midamble(s) and/or postamble for PDRCH transmission.
For other amble options, 
Source [3, Huawei] observed that
With accurate SFO estimation, Option 2 of D2R preamble + 4 midambles can achieve 10% BLER at SNR ~ 2.7dB but cannot achieve 1% BLER.
With up to 10% SFO, Option 3 cannot achieve 10% BLER.
With up to 10% SFO, Option 4 of D2R preamble+2 midambles+postamble achieves 10% BLER at SNR of ~0.25dB; But it cannot achieve 1% BLER. Option 4 of D2R preamble+3 or 4 midambles+postamble, achieves a 10% BLER at an SNR of around -0.2 dB, and achieves 1% BLER at SNR around 9dB or 8dB, respectively.
Source [5, CMCC] observed that with up to 10% SFO, Option 4 of D2R preamble combined with 1 to 4 midambles + postamble, achieves 10% BLER at SNR of 2.5 dB, 1 dB, 0.8 dB, or 0.5 dB, respectively, for a data rate of around 1 kbps.
Source [6, ZTE] observed that with up to 10% SFO, 
Option 3 can provide ~5.5 dB performance gain compared to option 2 of D2R preamble+1midamble for 10% BLER, with the same amble(s) overhead for ~1kbps data rate.
Option 2 of D2R preamble+1midamble cannot achieve 1% BLER for ~1kbps data rate.
Option 4 of the D2R preamble+1 or 2 midamble(s)+postamble, has similar performance, it can achieve a 10% BLER at SNR of -1dB and achieves a 1% BLER at SNR of 6dB and 5dB respectively for ~1kbps data rate.  
Source [8, Spreadtrum] observed that with up to 10% SFO, 
Option 3 of D2R preamble+ postamble cannot achieve 10% BLER for ~7kpbs.  
Option 4 of D2R preamble + 1 midamble + postamble can achieve 10% BLER and 1% BLER at SNR around -6dB and 0 dB, respectively for ~7kpbs data rate.
Source [9, vivo] observed that 
With up to 10% SFO, maintaining the same amble overhead, both Option 2 of D2R preamble+5 midambles and Option 4 of D2R preamble+4 midambles+1 postamble demonstrate similar performance, achieving 10% BLER at SNR ~0.2dB. While for 1% BLER, the SNR for Option 2 and Option 4 is ~ 9.2dB and 12.8dB, respectively for ~5.5kpbs data rate.   
With up to 1% SFO, maintaining the same amble overhead, both Option 2 of D2R preamble+5 midambles and Option 4 of D2R preamble+4 midambles+1postamble demonstrate similar performance, achieving 10% BLER at SNR around -1.2dB. While for 1% BLER, the SNR for Option 2 and Option 4 is ~7.8dB and 9.1dB, respectively for ~5.5kpbs data rate.   
Source [14, xiaomi] observed that with up to 10% SFO, Option 2 of D2R preamble+3 midambles and Option 4 of D2R preamble+3 midambles+postamble can achieve 10% BLER when the SNR is within the range of [15, 25] dB for 1.25 kbps data rate. 
Source [30, Qualcomm] observed that the required SFO estimation accuracy to achieve 1% and 10% BLER is much smaller than 10^3 ppm. To achieve the required accuracy,
For Option 2 of D2R preamble+X midamble(s) where midamble inserted per every certain number of PDRCH bits (e.g., 192 bits),
For SFO estimation using each amble for the subsequent PDRCH bits (e.g., 192 bits), with up to 10% SFO, more than 50 SFO hypotheses are necessary at the reader side and with up to 1% SFO, 6 SFO hypotheses are sufficient at the reader side.
For SFO estimation based on the time gap between preamble and midamble, with up to 10% SFO, 10 SFO hypotheses are used, but reader has to store the received samples and wait for the midamble to start SFO/channel/interference estimation, demodulation, decoding, etc.
For Option 3 of D2R preamble+postamble, SFO estimation is based on the time gap between preamble and postamble, with up to 10% device SFO, 10 SFO hypotheses are used for reader, but reader has to store the received samples and wait for the postamble that is after the end of PDRCH for any of SFO/channel/interference estimation, demodulation, decoding, etc
Source [7, Samsung] observes that the additional ambles i.e., midamble(s) and/or postamble introduces additional overhead and postamble may prevent pipelined processing of the reception.

For the synchronization and timing tracking of D2R transmission, 
Source [5, CMCC] report that with up to 10% SFO, option 1 is not sufficient for D2R reception since the residual SFO at reader side is larger than 1%. While with option 3, the reader can precisely search and detect the SFO with a residual SFO of 0.03% at -3dB SNR TDL-A channel.
Source [14, xiaomi] report that 
For packet size of 96bits, when the SNR is increased from -4dB to 20dB, the ratio of device residual SFO over 100ppm decreases to 6% for Option 2, 3 and 4, but remains at 95% for Option 1.
For packet size of 400bits, when the SNR is increased from -4dB to 20dB, the ratio of device residual SFO larger than 10ppm decreases to 5% for Option 2, 3, and 4, but is still 99.6% for Option 1.
Sources [9, vivo], [15, CATT] report that SFO estimation based on D2R preamble can achieve accurate estimation without additional ambles (midamble or postamble). 
Source [9, vivo][7 Samsung] observed that for non-coherent detection of PDRCH, the number of SFO hypotheses and the SNR needed for 10% and 1% BLER cannot significantly be reduced for option 2, 3 and 4 compared to the option 1. Moreover, the additional ambles i.e., midamble(s) and/or postamble introduces additional overhead and postamble may prevents pipelined processing of the reception.     
Source [15, CATT] observed that 
The coarse estimation of SFO based on the D2R preamble indicates that the SFO estimation error is less than 1% with a probability of 99.3%, and less than 0.1% with a probability of 49.9%.
The fine estimation of SFO based on the D2R preamble shows that the SFO estimation error is less than 1% with a probability of 99.5%, and less than 0.1% with a probability of 90.8%.
Reader/gNB can achieve a probability of not less than 99.5% for SFO estimation error below 1%, and 90.8% for SFO estimation error below 0.1% by receiving D2R preamble signals.
Source [30, Qualcomm] report that for D2R with coherent demodulation at reader, the reader needs to estimate the device clock frequency with the accuracy of 0.5% (5 * 10^3 ppm) or lower for a short message (e.g., 72 bits after CRC/coding) and 0.1% (10^3 ppm) or lower for a long message (e.g., 224 bits after CRC/coding). The source further reports that design of D2R amble(s) (e.g., overhead) and the correspondingly required number of SFO hypothesis for the estimation depend on the sampling clock accuracy that the device uses for D2R. 
Source [37, MediaTek] reports that transmitting 96-bit packet size with 16-bit CRC requires residue SFO after reader compensation to be 1000 ppm, and transmitting 1000-bit packet size with 16-bit CRC requires residue SFO after reader compensation to be 100 ppm.
Note: in the observations above where coherent detection is used, sources that evaluated option 3 and option 4 assumed that the postamble is used at least for time/frequency tracking and for channel estimation.

Agreement
Following observations on R2D clock-acquisition part are captured in TR 38.769:
On impact/restriction of M values for the clock-acquisition part
9 sources [TCL, Nokia, Huawei, CMCC, ZTE, Apple, CATT, Mediatek, Qualcomm] provided observations on the impact/restriction of M values for the clock-acquisition part design requirements:
1 source [Nokia] observed that increasing value of M, while retaining the same transmission duration, improves the auto-/cross- correlation properties of the sequence due to increase in sequence length and use of   provides better timing estimation accuracy even in the presence of SFO as the sequence length spans only over a shorter duration. 
2 sources [TCL, Huawei] observed for option 1 of the clock-acquisition part design that no restriction is required to be placed on the M values. Furthermore, 1 source [Huawei] observed that the same 2 ON-OFF voltage (with the same duration) satisfies the FDR performance metric of less than 1% for different M values, e.g., M = 2, 6 and 24, where FDR is the False detection ratio (FDR), i.e. incorrectly calculating M, is the performance metric.
1 source [CMCC] observed that pattern of the clock-acquisition part is related to M chips per OFDM symbol and when M is small, the clock-acquisition part may cross multiple OFDM symbols, and the CP insertion may degrade the timing acquisition performance.
1 source [ZTE] observed that with option 2, the duration of the clock-acquisition part remains consistent across all M values, at least three OFDM symbols maybe required for clock-acquisition part and it maybe not as efficient as option1
1 source [Apple] observed that among the two options studied for the clock-acquisition part, option 2 provides increased robustness, especially in case of large value of M, when compared to option 1 and potentially increase the detection performance of the clock-acquisition part. 
1 source [CATT] observed that if the chip duration is variable based on the M value used for OOK-4 waveform, the detection performance would be limited by the received SINR of the CAP with clear transition of the rising and falling edges.
1 source [Qualcomm] observed that the option 1 with M>1 has shorter duration of clock acquisition part than M=1 and worse timing acquisition accuracy. At least part of PRDCH following the clock acquisition part may need to be used to improve the timing acquisition. Furthermore, the larger M (e.g., M>4) with small chip duration is more sensitive to the SFO accuracy and the restriction of M for the clock acquisition part may be needed.
1 source [Mediatek] further observed that different M values may impact the chip accuracy obtained by the clock acquisition part.

On impact of CP insertion/handling on the clock-acquisition part
10 sources [TCL, CMCC, ZTE, Samsung, Vivo, CATT, NTT Docomo, Qualcomm, Mediatek, Spreadtrum] observed that the CP insertion/handling may impact the design requirements of the clock-acquisition part:
1 source [CMCC] further observed that when the clock-acquisition part occupies more than one OFDM symbol, ON-OFF state transition around CP can avoid the error rising or falling edges due to the CP insertion.
1 source [ZTE] further observed that to mitigate the impact of the CP in the clock-acquisition part for large M values, it can reuse the CP handling method for PRDCH 
1 source [Samsung] further observed that CP insertion/handling on the clock-acquisition part can cause false rising/falling transition and, therefore, the clock acquisition part should be designed such that it does not incur a false rising or falling edges due to CP insertion when CP-OFDM is used for OOK signal generation.
1 source [vivo] further observed that CP insertion/handling on the clock acquisition part will impact the chip duration estimation accuracy. It is further observed that for CP handling, device may not be able to count the clock and estimate OFDM symbol duration accurately until the clock acquisition part if the start indicator only includes a single ON-OFF transmission. 
1 source [CATT] further observed that the SER will be degraded due to uneven chip interval when the CP is inserted within an OFDM symbol, where SER refers to the number of samples which is mismatched for comparing to the total number of samples in a chip.
1 source [NTT Docomo] further observed if CP insertion would cause false rising/falling edges, accuracy of timing acquisition may be impacted.   
1 source [Mediatek] further observed that the issues of chip extension, false raising/falling transition, and additional raising/falling transition caused by CP insertion/handling considering different M values will impact the chip accuracy obtained by the clock acquisition part.
1 source [Spreadtrum] further observed that the design of clock acquisition part should consider that CP insertion does not cause a false rising or falling edges and does not cause different length of multiple high / low voltages within the clock acquisition part when the clock acquisition spans multiple OFDM symbols.
1 source [Huawei] observed CP insertion/handling may not impact the design requirements of the clock-acquisition part

Agreement
For the D2R preamble design, following aspects have been studied and can be captured in the TR 38.769:
Autocorrelation Property
10 sources [Nokia, Huawei, CMCC, Xiaomi, CATT, Oppo, Ericsson, NTT Docomo, Qualcomm, ZTE] observed that the signal should have good autocorrelation properties for accurate peak detection based on the signal correlation at the reader 
Cross-correlation Property
7 sources [Nokia, CMCC, Oppo, Ericsson, Qualcomm, ZTE, CATT] observed that the signal should have good cross-correlation properties if multiple D2R preamble sequences are considered (e.g. for multiple access schemes (if supported) for D2R transmissions). 
Line coding
1 source [Nokia] observed that line coding may impact the autocorrelation property of the sequence. 
1 source [Huawei] observed that for D2R preamble, to apply backscattering, line coding can help improve the detection performance based on shifting the D2R signal’s frequency location away from the carrier wave
Sequence Types (not limited to below types only)
M-sequence
3 sources [Nokia, Vivo, Xiaomi] observed that m-sequence can be considered for D2R preamble mainly owing to good correlation properties.  
Golay sequence
4 sources [CMCC, Vivo, Xiaomi, Samsung] observed that Golay sequence can be considered for D2R preamble mainly owing to good correlation properties and availability of large number of distinct sequences and complementary pairs.  
Walsh sequence
1 source [Oppo] observed that Walsh sequence can be considered as a candidate for D2R preamble thanks to its good auto/cross-correlation property and flexible length
General Observations
1 source [Huawei] observed can achieve 0.97% residual SFO with 98% probability under -2.5dB SNR and 0.1% MDR with [-1/8, 1/8] chip timing error with 99.05% probability under -2.5dB SNR with D2R preamble including 2-parts with clock-like sampling frequency signal and timing-acquisition signal, having 32-length ‘1’ sequence (encoded to 64-chip Manchester code) and 32-length sequence (encoded to 64-chip Manchester code), respectively.
4 sources [TCL, CMCC, ZTE, Vivo] observed that for D2R preamble with binary signal, the timing synchronization performance is highly related to the sequence length of the preamble. Furthermore, 1 source [CMCC] observed that to achieve a BLER performance at 10%, the timing synchronization error should be less than 10%. Furthermore, 1 source [ZTE] observed that the channel estimation performance is also highly related to sequence length. 1 source [ZTE] observed that using a 32 bits preamble provides ~8 dB, ~5 dB performance gain than using 8 bits, 16 bits preamble, respectively. And using a 64 bits preamble provides ~2.5dB performance gain than using a 32 bits preamble.
1 source [Ericsson] observed that for D2R preamble with binary signal, normalized SFO estimation error of less than 10% can be achieved with a training sequence length 64 or longer. The simulated D2R preamble consisting of a Golay complementary pair can tolerate SFO up to 1% (AWGN) with up to 1 dB loss in performance for a sufficiently long preamble sequence length (32 or greater).

Agreement
For determining the end of PRDCH at the device, following two options are studied and captured in the TR 38.769:
Option 1: TBS information (via implicit/explicit L1 R2D control information)
Option 2: Postamble (at the end of PRDCH) 
14 sources [Nokia, Huawei, ZTE, CMCC, Samsung, Ericsson, Oppo, LGE, Qualcomm, Spreadtrum, Mediatek, Cewit, Ericsson, vivo] provided following observations on the above two options for determining the end of PRDCH:
3 sources [Nokia, Huawei, ZTE] observed that option 2 provide two benefits, namely, the variable payload length and to provide timing acquisition before the subsequent transmission of either PDRCH or PRDCH, thus improving the detectability at both reader and the device, respectively. Furthermore, 1 source [Huawei] observed that R2D postamble indicates the TBS with high efficiency for small packets by avoiding a large padding overhead, unlike option 1, which may require devices to perform blind detection of different PRDCH formats (if supported) and the overhead caused by the inclusion of a R2D postamble does not exceed 20% for even the smallest of message sizes and may be less than the signaling overhead caused by using a dedicated TBS indicator
1 source [CMCC] observed for option 2, that for small payload size with only a few bits, the presence of long postamble generates large resource overhead, while for large payload size with more bits, the resource overhead of postamble is smaller.
1 source [vivo] observed for option 2, that for small payload size with only a few bits, the presence of long postamble generates large resource overhead.
1 source [Samsung] observed option 2 is not strictly required, however, given the possible clock drift at a device, it may be still beneficial to also attach postamble at least for the determination of the end of PRDCH at a device. 
3 sources [Oppo, Spreadtrum, CEWiT] observed that with option 2, the false detection may be higher for shorter postamble. Source [OPPO[ observed that in contrast to option 2, it is more reliable and efficient to indicate TBS with control information in option 1
2 sources [LGE, vivo] observed that if a message type or a command ID is included in L1 control information and implicitly indicates a known size of a fixed TB, then there is no need for either option 1 or option 2
2 sources [Qualcomm, vivo] observed that option 1 has the advantages of avoiding blind detection of postamble and providing the power saving for non-target devices to skip the R2D detection.
1 source [MediaTek] observed that option 1 is feasible for the device to avoid the unnecessary reception of a TB with a specific size and thus enable power saving, e.g., when the TB has a size exceeding the allowance of the device remaining power.
1 source [Ericsson] observed option 2 is not strictly required if the end of PRDCH can be explicitly indicated by R2D control information, and it is subject to the miss-detection rate. It may be beneficial if a PRDCH postamble can serve as an additional timing acquisition signal prior to a PDRCH transmission.

Agreement
For D2R scheduling, midamble (if supported) related information can be explicitly/implicitly indicated via corresponding PRDCH.

Agreement
Following observations on R2D clock-acquisition part are additionally captured in TR 38.769:
On purpose of SFO estimation/correction based on the clock-acquisition part
3 sources [Nokia, CATT, Qualcomm] provided observations on the applicability of clock-acquisition part for frequency synchronization:
1 source [Nokia] observed that the length of preamble sequence may need to consider also the robustness against SFO
1 source [CATT] observed that device 2a/2b may require higher synchronization accuracy for signal transmission or backscattering and therefore, the design of CAP may be required to accommodate the requirement of additional frequency synchronization and clock calibration for Device 2a/2b. 
1 source [Qualcomm] further observed for Option 1, as the CAP duration with high M is decreased, only CAP may not be sufficient for SFO correction and for Option 2, as the CAP duration is fixed and independent from M, the CAP with long enough duration can support SFO correction.
On purpose of CFO estimation/correction based on the clock-acquisition part
2 sources [Ericsson, Qualcomm] provided observations on the applicability of clock-acquisition part for frequency synchronization:
1 source [Ericsson] observes that the clock-acquisition part can be utilized to solve the frequency synchronization problem without impacting the time-domain sequence, for example by transmitting in some frequency resources and it can be a harmonized solution for both chip duration indication and device frequency synchronization. However, it is further observed that if the time interval between an R2D transmission and the corresponding D2R transmission and if the device loses the timing obtained from the R2D timing acquisition signal due to timing drift at the time for the D2R transmission, then an additional synchronization signal is needed
1 source [Qualcomm] further observed for Option 1, as the CAP duration with high M is decreased, only CAP may not be sufficient for CFO correction and for Option 2, as the CAP duration is fixed and independent from M, the CAP with long enough duration can support CFO correction.


WI Phase: RAN1 Agreements (relevant for R2D/D2R signals including timing acquisition and synchronization
RAN1#120 (Athens, Greece, Feb 17th – 21st, 2025)
SIP related Agreements
Agreement
For the SIP of R-TAS, for providing the start of the R2D transmission, one single design based on Option 1 is supported and further down-selection to be done among Alt 1 and Alt 2 :
Option 1: ON-OFF transmission with following alternatives:
Alt 1: A single ON-OFF transmission with pre-defined duration for each of the ON-OFF, where ON and OFF may have same or different durations
Continue discussion to clarify the duration of each of the ON and OFF
Continue discussion to list the different candidate proposals under Alt1
Alt 2: A multi-ON-OFF transmission with pre-defined duration for each of the ON(s)-OFF(s), where different ON and different OFF may have same or different durations and different parts may have same or different duration
Continue discussion to clarify the duration of each of the ON and OFF
Continue discussion to list the different candidate proposals under Alt2
Only a single fixed value for entire duration of SIP of R-TAS is supported, which is independent of the value of “M” used in CAP and PRDCH
Note: Specific design and duration for SIP of R-TAS are further discussed, and companies are encouraged to evaluate the designs in terms of target MDR of [10%] for a FAR up to [1%] and at least following assumptions are used:
MDR refers to the probability that SIP is not detected when it was actually transmitted
FAR probability that the receiver incorrectly detects SIP when SIP was not transmitted
Energy/edge detection-based method is the baseline assumption for evaluation purpose
Continue discussion on necessary details for simulation assumptions

Agreement
For the SIP of R-TAS, down-select among the following candidates:
Alt 1 (Single ON-OFF transmission)
Alt 1-1: ON followed by OFF with same duration for both
Alt 1-2: ON followed by OFF with a duration ratio of 1:[2,3]
Alt 1-3: ON followed by OFF with a duration ratio of [2,3]:1
Alt 2 (Multi-ON-OFF transmission)
Alt 2-1: A number of repetition instances of Alt 1-1 or Alt 1-2 or Alt 1-3
Alt 2-2: ON-OFF-ON (duration of ON and OFF can be different)
Alt 2-3: OFF-ON-OFF (duration of ON and OFF can be different)
Alt 2-4: Combination of single instance of Alt 1-1 and single instance of Alt 1-2
For the evaluation purpose, for both options, candidate values related to duration are considered:
Entire duration of SIP: 1/2 OFDM symbol duration or 1 OFDM symbol duration (including clarifying whether OFDM symbol duration includes CP); additional durations can be considered and reported by companies with justification 
Companies to report the exact duration(s) for ON or OFF
Companies are encouraged to report at least the following details for the evaluations:
Baseline assumption is that RF transmission is not present; companies can report other consideration
For FAR calculation, whether noise and/or PRDCH transmission is considered
Details on threshold detection method including whether/how threshold detection training is used based on the proposed design alternative or not
BW assumptions for RF-ED and BB-LPF
Target MDR of up to 1% for FAR of up to [1%, 10%]

CAP related Agreements
Agreement
For the CAP of R-TAS, the starting chip has a different voltage level compared to the end of the SIP of R-TAS.

Agreement
For the design of the CAP of R-TAS, at least 2 transition edges in same direction are included, i.e. at least two transitions from “OFF” chip to “ON” chip or two transitions from “ON” chip to “OFF” chip.

Agreement
For the CAP of R-TAS:
Candidate values for maximum duration of CAP to be further down-selected to one value from : 1.5 OFDM symbol duration, 2 OFDM symbol duration, 3 OFDM symbol duration
For option 1 for CAP of R-TAS from TR 38.769, maximum duration is applicable to minimum value of M to be supported, and the CAP duration becomes shorter with increasing value of M
FFS: whether the number of ON/OFF transmissions in the CAP is fixed or not fixed
For option 2 for CAP of R-TAS from TR 38.769, maximum duration is the only (constant) duration that is applicable for all the M values to be supported
Down-selection between option 1 and option 2 for CAP of R-TAS from TR 38.769 by RAN1#120-bis
FFS: Values of M to be supported


R2D Midamble related Agreement
Agreement
R2D transmission does not include a midamble.

D2R X-amables related Agreement
Agreement
For D2R preamble design, the functionalities of timing acquisition, SFO estimation/time tracking and channel estimation should be supported
For D2R midamble design, the functionalities of SFO estimation/time tracking and channel estimation should be supported
D2R midamble can be transmitted at the end of the PDRCH transmission. If it is at the end, it is not designed for being used for indicating the end of PDRCH transmission
FFS: condition(s) and/or indication where the D2R midamble is present or not

Agreement
For D2R x-ambles:
Following is considered as the types for base sequence and to be further down-selected:
Option 1: M-sequence 
Option 2: Golay sequence
Note: Above doesn’t preclude an additional part for preamble, e.g. with ON and/or OFF transmission, if needed/supported
FFS: Whether/what multiple sequences (using same base sequence type) are supported
Note: This in no way implies that there is going to be CDMA between D2R x-ambles
For evaluation purpose, companies are encouraged to consider following:
Performance at least in terms of autocorrelation/cross-correlation property, SFO estimation/Timing accuracy, SNR for target PDRCH BLER of [1%, 10%]
Report presence and time-domain resource(s) x-ambles
Report sequence type(s) and length(s) for x-ambles
Following format can be considered for reporting the evaluation results


RAN1#120bis (Wuhan, China, April 7th – 11th, 2025)
Agreement
For D2R midamble, for determining the presence and location of midamble(s) at the device:
Reader explicitly indicates the same interval between consecutive midambles, and between the preamble and the first midamble, via R2D control information
FFS: details of signalling
FFS: whether the reader can explicitly indicate with one bit whether a midamble is additionally present at the end
Note: This does not preclude the support of having no midamble present in the D2R transmission

Agreement
For the pattern of SIP of R-TAS, only the following 2 alternatives are considered for further down-selection:
Alt 1-2: ON-OFF with a ratio of 1:3 and with following total SIP duration to be further down-selected:
Option 1: 0.5 OFDM symbol duration
Option 2: 1 OFDM symbol duration
Alt 2-4: ON-OFF-ON-OFF with a ratio of 1:1:1:3 and with following total SIP duration to be further down-selected:
Option 1: 0.5 OFDM symbol duration
Option 2: 1 OFDM symbol duration

Agreement
For D2R preamble/midamble, base sequence is generated from m-sequence, where the length of the sequence is 
Value(s) of n
Long preamble/midamble is generated based on n = 5
Working assumption: Short preamble/midamble is generated based on n=3 
Only 1-part preamble/midamble are supported for D2R
Preamble immediately precedes the PDRCH without any gap
Both long and short preamble and midamble are supported based on the working assumption on n
when midamble is present at least the following cases are supported and reader explicitly indicates one of the following cases for PDRCH:
Short preamble and short midamble 
Long preamble and long midamble 
Note: the case of short preamble and long midamble will not be supported
When midamble is not present the reader explicitly indicates short or long preamble for PDRCH

Agreement
For CAP of R-TAS, following is adopted:
Option 1 for CAP of R-TAS from TR 38.769 is adopted where the CAP duration becomes proportionally shorter with increasing value of M, i.e. if for , duration is  OFDM symbol long, then for , duration is  OFDM symbol long
Note: Duration without CP insertion is considered above, with CP insertion, the total duration may not be exactly proportional
Only following two alternatives for CAP pattern are considered for further down-selection to one alternative:
Alt 1: ON-OFF-ON-OFF
Alt 2: ON-OFF-ON

Agreement
For indicating the interval between consecutive midambles, and between the preamble and the first midamble, via R2D control information, following is adopted:
Unit of interval is number of bits after FEC (if FEC is applied) and repetition (if repetition is applied)
FFS: the candidate values in terms of the unit of interval

TDoc file conclusion not found

3GPP TSG RAN WG1 #120bis	R1-2502664
Wuhan, China, April 7th – 11th, 2025

Agenda Item:	9.4.3
Source:	Fraunhofer HHI, Fraunhofer IIS 
Title:	Discussion on timing and synchronization aspects
Document for:	Discussion

Conclusion
In this contribution we discussed the design aspects of the timing and synchronization signals for A-IoT. Based on the discussion in the previous sections we made the following observations and proposals:

Observation 1. The SIP design presents a trade-off between power efficiency and detection robustness. Shorter ON durations reduce power consumption, whereas longer ON durations enhance resilience against interference
Observation 2. If an energy detection mechanism is used for SIP detection, Type1 devices may struggle to detect the SIP if its duration is not long enough

Proposal 1. For the R2D start indication signal, a multi ON-OFF approach (Alt 2) is supported.
Proposal 2. For the R2D SIP, support Alt 2 (multi ON-OFF transmission) and down-select from Alt 2-1 and Alt 2-2.
Proposal 3. For the R2D SIP, the time duration used for the ON and OFF periods should be different, and distinguishable from the rest of the frame.
Proposal 4. Different M values within the fixed SIP duration should be leveraged to provide a filtering mechanism, reducing energy consumption for non-targeted devices.
Proposal 5. For R2D CAP, support option 1 with a fixed number of ON/OFF transmissions in order to enhance synchronization reliability across different device types.
Proposal 6. For D2R preamble, support defining a short and long design.

3GPP TSG RAN WG1 #120-bis                               R1-2502673
Wuhan, China, February 7th - 11th, 2025

Source:	Sharp
Title:	Discussion on timing acquisition and synchronization
Agenda Item:	9.4.3
Document for:	Discussion and Decision

Conclusion
In this contribution, we discuss a few aspects related to timing acquisition and synchronization in A-IoT, and make the following observations and proposals.
For R2D, additional time units other than R2D chip lengths determined by the supported values of “M”s are undesirable.
For R2D SIP, if the duration is not based on one of the supported values of “M”, it is unclear how the restriction of minimum R2D bandwidth is applied.
In R2D clock-acquisition part (CAP) Option 2, the duration of the R2D CAP which is constant for different M values based on repetition may be too long (in terms of detection performance) for M>1.
For R2D SIP, each ON or OFF has an integer multiple of R2D chip lengths based on a reference value of “M”, i.e. each ON or OFF consists of one or more R2D chips based on that value of “M”.
Entire duration of R2D SIP is 1/2 OFDM symbol duration (excluding CP).
For R2D clock-acquisition part (CAP), adopt Option 1 (i.e. duration of the R2D CAP is variable for different M values, where the duration becomes shorter with increasing value of M).
The number of ON/OFF transmissions in the CAP is a fixed value.
For D2R X-ambles, support multiple sequences (using same base sequence type).

3GPP TSG RAN WG1 Meeting #120bis	R1-2502706
Wuhan, China, April 7th – 11st, 2025

Source:	MediaTek Inc.
Title:	A-IoT - Timing Acquisition and Synchronization 
Agenda item:	9.4.3
Document for:	Discussion and decision

Conclusion
In this contribution, we have the following proposals:
Observation 1: Maximum duration of CAP as 3 OFDM symbol duration is more flexible to accommodate more potential M values and chip numbers within a CAP.
Observation 2: Regarding CAP design, option 1 with a not fixed number of ON/OFF transmissions in the CAP is more flexible to balance the overhead and performance.
Observation 3: The max TB size up to 1000 bits does not necessarily mean there is 1000 possibilities of TB size. The overhead of option 1 is controllable via supporting a limit number of TB sizes, e.g., a limit typical commands for PRDCH as RFID, or a relative coarse TB size granularity.
Observation 4: Option 1 has the benefit to avoid blind detection at the device side compared to option 2.
Proposal 1: For the CAP of R-TAS, the maximum duration of CAP is 3 OFDM symbol duration and option 1 for CAP of R-TAS from TR 38.769 with a not fixed number of ON/OFF transmission in the CAP is supported.
Proposal 2: To determine or derive the end of PRDCH transmission, support TBS information via implicit/explicit R2D control information.

3GPP TSG RAN WG1 #120bis			R1-2502768 
Wuhan, China, April 7th – 11st, 2025

Source:	NTT DOCOMO, INC.
Title:	Discussion on timing acquisition and synchronization for Ambient IoT
Agenda Item:	9.4.3
Document for: 	Discussion and Decision

Conclusion
In this contribution, we discussed timing acquisition and synchronization aspects of A-IoT. Based on the discussion, we made the following proposals.
Proposal 1:
Support the following alternative for SIP in R-TAS.
Alt 2-1: A number of repetition instances of Alt.1-2/Alt.1-3, where Alt.1-2/Alt.1-3 is ON followed by OFF with a duration ratio of 1:3/3:1.
Alt 2-4: Combination of single instance of Alt.1-1 and single instance of Alt.1-2, where Alt.1-2 is ON followed by OFF with a duration ratio of 1:3.
Proposal 2:
Duration of chips in CAP is uniform after CP insertion. 
Proposal 3:
Chip duration of R2D control is determined from CAP, where candidate M values of R2D control can be selected from {1, 2, 4, 6}.
Chip duration of R2D data is indicated in R2D control.
Proposal 4:
The design of R2D clock acquisition part should targetthe performance which can meet the accuracy after clock sync / calibration at device side as 104 ppm.
Proposal 5: 
Start and/or end of R-SIP/CAP is aligned with OFDM symbol boundary.
Proposal 6:
Device determines the end of R2D transmission based on violation of Manchester coding rule corresponding to the M value of the received PRDCH. 
The violation rule is specified, e.g., device detects a high/low voltage with duration >= 3 chips.
FFS distinguishing the end of R2D transmission from R-TAS-SIP.
Proposal 7:
Support multiple sequences with same length and with good cross correlation property.
FFS how to determine which sequence is used for a D2R transmission.
Proposal 8:
Multiple sequences with different length are not supported.
Proposal 9:
D2R midamble can be transmitted at least in D2R Msg.3 and data transfer after random access.
Proposal 10:
Support explicit indication of presence of D2R midamble and gap between two D2R x-ambles in corresponding R2D. 

3GPP TSG RAN WG1 Meeting #120bis    	            R1-2502841
Wuhan, China, April 7th – 11st, 2025

Source:	Qualcomm Incorporated
Title:	Timing acquisition and synchronization
Agenda Item:	9.4.3
Document for: 	Discussion and Decision

Conclusion
In this contribution we have following proposals and observations.

Start indicator part of R2D timing acquisition signal
Observation 1:
For an SIP with a single ON-OFF transmission, the SIP detection is based on the OFF period of the SIP
A device considers the SIP is detected if the number of 0s from the comparator output matches with the OFF period of the SIP

Observation 2:
For single ON-OFF transmission,
Longer ON period before OFF period offers lower miss-detection probability due to better comparator threshold training.
Longer OFF period offers lower false-alarm probability under noise input.
OFF period duration should be deviated from any OFF periods expected during CAP/PRDCH OOK as much as possible if lower false-alarm probability under CAP/PRDCH envelope input is desirable.

Observation 3:
Without digital correlator, multiple ON-OFF transmission has no/less advantage compared to single ON-OFF transmission for a given SIP duration

Proposal 1:
SIP is [1000] (M=4) on one OFDM symbol

Clock acquisition part of R2D timing acquisition signal
Observation 4:
3 OOK chip duration in CAP can reliably indicate the M value if CAP M values are limited to {1, 2, 4, 6, 12} 

Proposal 2:
Confirm to design clock-acquisition part such that following functionalities are enabled:
Identify the value of M used for the PRDCH, at least if CP handling Alt M1-1/1-2 is agreed
Determine the OOK chip duration
The determined OOK chip duration is a reference for D2R transmission

Proposal 3:
For clock-acquisition part for a device to determine the OOK chip duration, adopt following:
Option 2: Duration of the clock-acquisition part is constant for different M values based on repetition, i.e. repetition factor is increased with increasing value of M to keep the duration constant.
The CAP has 3 OFDM symbols

Proposal 4:
For CAP Option 2 over 3 OFDM symbols, support M values of up to 12
Alt.1: Define mapping between M value of CAP and M value of PRDCH
E.g., {M for CAP, M for PRDCH} = {1, 2}, {2, 4}, {4, 8}, {6, 12}, {12, 24}
Alt.2: Support two-parts PRDCH
First part of PRDCH uses the same M value as CAP
Second part of PRDCH uses the M value indicated by the first part of PRDCH
Regarding CP handling for CAP Option 2,
For CAP M = 1, CP handling is not necessary
For CAP M = 2, 4, 6, CAP pattern is cyclically shifted in time such that CP does not create fake rising/falling edges or fluctuate OOK chip duration in the middle of the CAP
The first OOK chip and the last OOK chip of an OFDM symbol of a CAP are ON over 0.5 OOK chip duration for a given CAP M
For CAP M = 12, CAP pattern after CP insertion over 3 OFDM symbol is [1010101010101 0101010101010 10101010101]
An CP duration is handled as an OOK chip of the CAP

R2D postamble
Proposal 5:
Not to specify physical layer R2D postamble
A-IoT device may consider an R2D transmission ends when it does not detect Manchester coding codeword in the R2D transmission

D2R preamble and midamble
Observation 6:
With the device SFO for D2R transmission in a range [-10%, +10%],
For a PDRCH for 16 bits + CRC-6 encoded by TBCC R = 1/3 (66 bits in total), BLER 1% is achievable with preamble only
For a PDRCH for 96 bits + CRC-16 encoded by TBCC R = 1/3 (336 bits in total), BLER 1% is not achievable with preamble only

Observation 7:
For a PDRCH carrying a data of 16 bits + CRC-6 encoded by TBCC R = 1/3 (66 bits in total) with “preamble only”, with the device SFO for D2R transmission in a range [-10%, +10%],
If the sequence has 7 or 8 bits, BLER 1% is not achievable
If the sequence has 15 or 16 bits, autocorrelation properties of sequences have large impact on BLER performance
Golay sequence offers 2-3 dB performance gain compared to M and Gold sequences
If the sequence has 31 or 32 bits, impact of autocorrelation properties is marginal.
All the three sequence types offer quite similar BLER performances
Sequence length of 31 or 32 bits offers 2 – 8 dB gain compared to 15 or 16 bits for BLER 1%

Observation 8:
For a PDRCH carrying a data of 16 bits + CRC-6 encoded by TBCC R = 1/3 (66 bits in total) with “preamble + midamble”, with the device SFO for D2R transmission in a range [-10%, +10%],
The location of the midamble for the PDRCH does not impact on BLER performances
If the sequence is Gold sequence and has 7 or 8 bits, BLER 1% is not achievable
If the sequence has 15 or 16 bits, autocorrelation properties of sequences have large impact on BLER performance
If the sequence has 31 or 32 bits, impact of autocorrelation properties is marginal.
All the three sequence types offer quite similar BLER performances
Sequence length of 31 or 32 bits offers 2 – 5 dB gain compared to 15 or 16 bits for BLER 1%
For a PDRCH carrying a data of 16 bits + CRC-6 encoded by TBCC R = 1/3 (66 bits in total), with the device SFO for D2R transmission in a range [-10%, +10%],
“preamble + midamble” with 15/16 bits sequence length and “preamble only” with 31/32 bits length require almost same SNR for achieving BLER 1%
“preamble + midamble” with 31/32 bits sequence length achieves 2-4 dB performance gain compared to “preamble only” with 31/32 bits sequence length

Observation 9:
For a PDRCH carrying a data of 96 bits + CRC-16 encoded by TBCC R = 1/3 (336 bits in total) with “preamble + midamble”, with the device SFO for D2R transmission in a range [-10%, +10%],
The location of the midamble for the PDRCH does not impact on BLER performances
If the sequence is Gold sequence and has 7 or 8 bits, BLER 1% is not achievable
If the sequence has 15 or 16 bits, autocorrelation properties of sequences have large impact on BLER performance
If the sequence has 31 or 32 bits, impact of autocorrelation properties is marginal.
All the three sequence types offer quite similar BLER performances
Sequence length of 31 or 32 bits offers non-negligible performance gain compared to 15 or 16 bits

Takeaway:
With the device SFO for D2R transmission in a range [-10%, +10%],
Sequence length of 31/32 bits offers always better BLER performances than the shorter sequence lengths
For a PDRCH carrying a data of 16 bits + CRC-6 encoded by TBCC R = 1/3 (66 bits in total), both “preamble-only (31/31 bits)” and “preamble (31/32 bits) + midamble (31/32 bits)” works
“preamble-only (31/31 bits)” offers lower overhead (higher efficiency) at the cost of higher required SNR
“preamble (31/32 bits) + midamble (31/32 bits)” offers lower operating SNR at the cost of higher overhead
For a PDRCH carrying a data of 96 bits + CRC-6 encoded by TBCC R = 1/3 (66 bits in total), “preamble (31/32 bits) + midamble (31/32 bits)” is necessary
Midamble (31/32 bits) can be located at around the middle or at around the end of the PDRCH

Proposal 6:
If the device SFO for D2R transmission is in the range [-10%, +10%], adopt 31-bit Gold sequence for D2R preamble/midamble

Proposal 7:
If the device SFO for D2R transmission is in the range [-10%, +10%],
For a PDRCH for 16 bits + CRC-6 encoded by TBCC R = 1/3 (66 bits in total), support R2D control to indicate whether the D2R has “preamble only” or “preamble + midamble”
Midamble, if present, is located at around the end of the PDRCH
For a PDRCH for 96 bits + CRC-16 encoded by TBCC R = 1/3 (336 bits in total), support at least “preamble + midamble” with the midamble located at around the middle of the PDRCH
FFS: R2D control to indicate the location of the midamble or the number of midamble(s)

Device SFO [-10%, +10%] or [-1%, +1%] for D2R transmission
Observation 10:
Clock-acquisition part Option 2 can reduce device SFO for D2R transmissions to be within [-1%, +1%] and saves the total overhead of A-IoT communications

3GPP TSG RAN WG1 #120bis	R1-2502896
Wuhan, China, April 7th – 11th, 2025

Agenda Item:	9.4.3
Source:	Fraunhofer HHI, Fraunhofer IIS 
Title:	Timing acquisition and synchronization aspects for Ambient-IoT
Document for:	Discussion

Conclusion
In this contribution we discussed the design aspects of the timing and synchronization signals for A-IoT. Based on the discussion in the previous sections we made the following observations and proposals:

Observation 1. If an energy detection mechanism is used for SIP detection, Type1 devices may struggle to detect the SIP if its duration is not long enough

Proposal 1. For the R2D start indication signal, a multi ON-OFF approach (Alt 2) is supported.
Proposal 2. For the R2D SIP, support Alt 2 (multi ON-OFF transmission) and down-select from the following:
Alt 2-1 with repetitions of Alt 1-2 or Alt 1-3, 
Alt 2-2.
Proposal 3. Different M values within the fixed SIP duration should be leveraged to provide a filtering mechanism, reducing energy consumption for non-targeted devices.
Proposal 4. For R2D CAP, support option 1 with a fixed number of ON/OFF transmissions in order to enhance synchronization reliability across different device types.
Observation 2: Although, the timing drift is corrected during CAP, the device still suffers from timing drift during PRDCH transmission. 

Proposal 5: Study reporting of timing drift information from the device to the reader.
Proposal 6. For D2R preamble, support defining a short and long design.

3GPP TSG RAN WG1 Meeting #120	bis                                                               R1-2502915
Wuhan, China, April 7th – 11th, 2025 

Source:	CEWiT
Title:	Discussion on timing acquisition and synchronization aspects for Ambient IoT
Agenda Item:	9.4.3
Document for:	Discussion and Decision

Conclusion
In this document we discussed on timing acquisition, synchronization, and necessity of preamble, midamble, and postamble for A-IoT in both R2D and D2R transmissions. The following proposals are made: 
Observation 1: For start indicator part design:
ON-OFF transmission based on energy/edge detection (option 1) is simple to implement and consumes less energy for detection.
Multi-ON-OFF transmission (option1-Alt 2) pattern can reduce the probability of mis detection compared to a single ON-OFF transmission (option 1-Alt 1)
ON-OFF sequence-based (option 2) design reduces the probability of mis detection but increases the complexity and power consumption of a device. 
Proposal 1: For Start indicator part design, multi-ON-OFF transmission based on energy/edge detection (option1-Alt 2) is supported.
Proposal 2: Multi-ON-OFF transmission Alt 2-1 with repetition of Alt 1-2 pattern (Alt 2-1: A number of repetition instances of Alt 1-2) for Start indicator part design is supported.
Proposal 3: Support Option 1 for CAP of R-TAS 
Duration of the clock-acquisition part is variable for different M values, i.e. the duration becomes shorter with increasing value of M.
Observation 2: Preamble should be differentiable from the PRDCH.
Proposal 4: Following options are supported for differentiating the preamble from the PRDCH: 
Preamble followed by a long low voltage or guard time
Preamble followed by a TR2D_R2D_min for PRDCH
Observation 3: Regarding the options to determine or derive the end of PRDCH transmission:
Option 1 is more robust, lower miss/false-detection probability compared to option 2
In option 1, no additional power consumption caused by the miss detection compared to option 2
Option 1 offers greater flexibility for scheduling and operations
Option 1 can be used to mitigate the postamble overhead
Option 2 increases power consumption and detection failure if a device miss/false-detects the postamble. 
In option 2, additional signal needs to be defined and to be detected by the device
Option 2 adds latency delay to the decoding
Proposal 5: To determine or derive the end of PRDCH transmission, Option 1 (L1 R2D control information) is supported.
Proposal 6: Support midamble for D2R transmission large packet size.
Proposal 7: For D2R midamble design, the following aspect should be specified:
Determining the position and length of the midamble
The pattern for midamble
How to determine the number of midambles
Proposal 8: Following options can be considered for determining the position and the length of the R2D midamble:
Predefined rule 
D2R scheduling information
Proposal 9: To determine or derive the end of PDRCH transmission, Option 2 (based on control information) is supported.

TDoc file unavailable

3GPP TSG RAN WG1 #120bis		        					     R1-2503123
Wuhan, China, April 7th – 11th, 2025

Agenda Item:	9.4.3
Source: 	Moderator (Apple)
Title:	Final FL Summary on timing acquisition & synchronization for Ambient IoT
Document for:	Discussion & Decision

Introduction & Work-Plan for RAN1#120bis
This document provides the feature lead summary on the offline discussions/inputs/proposals for AI 9.4.3 timing acquisition and synchronization signals for R2D and D2R for ambient IoT WI during RAN1#120bis. 

Guidance for RAN1#121 (last meeting for Rel-19 Ambient IoT WI)
Following 3 issues are remaining to be completed for this agenda:

For SIP of R-TAS, for down-selection between SIP Alt 1-2 and Alt 2-4, companies are encouraged to provide their evaluations considering the assumptions and performance metrics that have been agreed during RAN1#120 and RAN1#120bis

For R2D postamble related aspects, following FL proposal and corresponding potential specification impact, if any, can be considered for further discussions and decision on whether R2D postamble is specified or not. And if it is agreed to be specified, preferred pattern with justifications are encouraged

Proposal 4-1d 
Manchester coding rule violation corresponding to M value for PRDCH is applied by reader’s implementation that can be used by the device to determine the end of PRDCH transmission
There is no consensus to specify R2D postamble

For D2R preamble/midamble, companies are encouraged to provide their preferred sequence for length corresponding to n = 5 and n =3 and potentially including evaluations considering the assumptions and performance metrics that have been agreed during RAN1#120
Also, companies are encouraged to provide the midamble interval values accordingly 
Corresponding signaling details related to D2R preamble/midamble will also be handled under this agenda


Contact Information
Please consider providing your company name, your name and email address to be able to reach for any potential offline discussions/contact regarding AI 9.4.3 on timing acquisition and synchronization for ambient IoT.


Work Plan
Following is my plan for this meeting for this agenda for RAN1#120bis:

SIP of R-TAS
For SIP, at least down-select to not more than 2 alternatives from the agreed alternatives in RAN1#120 considering technical justification and evaluation analysis. Also agree on fixed single duration for the down-selected alternatives. 
CAP of R-TAS
For CAP, plan is to down-select between option 1 and option 2 (as agreed in last meeting) and then also limit the number of specific CAP pattern design and hopefully have not more than 2 patterns for further down-selection by end of this meeting
R2D Postamble aspects
For indicating the end of PRDCH transmission, at least down-select to whether R2D control information is used, or Manchester coding rule violation is used. 
D2R X-amble
For D2R x-amble, first priority is to agree on the sequence type and length(s). In addition, for midambles, agree on the method/signaling that allows the device to determine the number of midamble(s) and their location within the PDRCH.




SIP of R-TAS

[Closed] 1st Discussion Round 
For the SIP of R-TAS, following aspects have been discussed in the contributions following the agreements and discussion from RAN1#120:
Preferred alternative(s) for SIP design
Duration of SIP
For comparison among different alternatives, companies also provided simulations results/analysis to justify their preferred alternative(s). In addition, a few companies also discussed other aspects including M value/chip duration for the SIP, whether/what impact of CP insertion on SIP. In the following table, companies’ views on their preferred alternatives along with justification and performance analysis, if provided are summarized. 

Table 2-1: Summary of views on alternatives for SIP



FL Observations:
28 companies provided their views on SIP of R-TAS
11 companies provided evaluation in terms of target MDR and FAR with most of the companies considering MDR <= 1% for a target FAR <= 1%
For Alt 1-1, 3 companies provided their preference with ON-OFF ratio of 1:1 & total duration spanning 1 OFDM symbol and only 1 company [China Telecom] showed that it performs better than other alternatives while it satisfies the target MDR and FAR. On the other hand, 4 companies [Futurewei, Xiaomi, Samsung, Huawei] showed other alternatives that perform better than Alt 1-1.
For Alt 1-2, overall, 8 companies provided their preference for at least some variant of Alt 1-2 and 6 companies [Futurewei, ZTE, Spreadtrum, Oppo, Xiaomi, Qualcomm] showed that at least some variant of Alt 1-2 performs better than other alternatives including Alt 1-1, Alt 1-3, Alt 2-2, Alt 2-4. On the other hand, 4 companies [China Telecom, Samsung, CMCC, Huawei] showed other alternatives that performs better than Alt 1-2. Within Alt 1-2:
2 companies prefer ON-OFF with ratio of 1:3 & total duration spanning 0.5 OFDM symbol
5 companies prefer ON-OFF with ratio of 1:3 & total duration spanning 1 OFDM symbol
3 companies prefer ON-OFF with ratio of 1:2 & total duration spanning 1 OFDM symbol
For Alt 1-3, overall, 6 companies provided their preference for at least some variant of Alt 1-3 and only 1 company [Samsung] showed that at least some variant of Alt 1-3 performs better than other alternatives including Alt 1-1, Alt 1-2. On the other hand, 6 companies [China Telecom, ZTE, Futurewei, Oppo, Qualcomm, Huawei] showed other alternatives that performs better than Alt 1-3. Within Alt 1-3:
4 companies prefer ON-OFF with ratio of 2:1 & total duration spanning 1 OFDM symbol
5 companies prefer ON-OFF with ratio of 3:1 & total duration spanning 1 OFDM symbol
For Alt 2-1, 7 companies provided their preference with ON-OFF ratio of 1:1 & total duration spanning 1 OFDM symbol and no company showed that it performs better than other alternatives. On the other hand, 2 companies [China Telecom, Vivo] showed other alternatives that perform better than Alt 2-1.
For Alt 2-2, overall, 6 companies provided their preference for at least some variant of Alt 2-2 and only 1 company [Vivo] showed that at least some variant of Alt 2-2 performs better than other alternatives including Alt 2-1, Alt 2-4. On the other hand, 2 companies [China Telecom, ZTE] showed other alternatives that performs better than Alt 2-2. Within Alt 2-2:
1 company prefer ON-OFF-ON with ratio of 1:1:4 & total duration spanning 1 OFDM symbol
2 companies prefer ON-OFF-ON with ratio of 1:4:1 & total duration spanning 1 OFDM symbol
3 companies prefer total duration spanning 1 OFDM symbol without any specific consideration on ration of ON/OFF
For Alt 2-3, overall, 3 companies provided their preference for at least some variant of Alt 2-3 and no company showed that it performs better than other alternatives including. On the other hand, 2 companies [China Telecom, Huawei] showed other alternatives that performs better than Alt 2-3. Within Alt 2-3:
2 companies prefer OFF-ON-OFF with ratio of 1:1:4 & total duration spanning 1 OFDM symbol
1 company prefer OFF-ON-OFF with ratio of 1:4:1 & total duration spanning 1 OFDM symbol
1 company prefer OFF-ON-OFF with total duration spanning 1 OFDM symbol
For Alt 2-4, 4 companies provided their preference with ON-OFF-ON-OFF ratio of 1:1:1:3 & total duration spanning 0.5 OFDM symbol and 2 companies [CMCC, Huawei] showed that it performs better than other alternatives. On the other hand, 2 companies [ZTE, Vivo] showed other alternatives that perform better than Alt 2-4.
A few companies discussed M value for SIP and whether/how the selection of M may impact the SIP detection considering CP insertion. However, based on discussions during RAN1#120, we considered discussing in terms of OFDM symbol duration and avoid reference M value. Therefore, with that understating, we should not try to reopen the discussion of reference M value but rather have discussion in terms of OFDM symbol duration. Eventually, it will be up to the specification editor on how to capture the SIP duration and corresponding ON(s) and OFF(s). Furthermore, at least 1 company raised the issue that for chip duration corresponding to M < 6, minimum PRB of 2 will be needed. However, another company clarified that we don’t need to associate the mapping of M to minimum number of PRBs for SIP and device should be able to adjust its LPF between SIP reception and follow-up R2D reception. 
Another aspect that a few discussed is whether/what impact CP might have on SIP detection. One company pointed out that chip duration corresponding to higher value of M map impact SIP detection, while 3 companies pointed out that CP impact is negligible on SIP and should not be considered for specific optimization, especially considering that all companies considered SIP duration of not more than 1 OFDM symbol

Based on above observations from companies’ contributions, as a starting point, FL proposal would be to down-select to Alt 1-2 and Alt 2-4. Main rationale behind these two alternatives is not just majority preference but evaluations showing that these two alternatives at least perform better within alternative set 1 and set 2, respectively. Further focused discussion on Alt 1-2 and Alt 2-4 can be further done, if we down-select these 2 options. Therefore, accordingly Proposal 2-1 is provided below:

(Updated) Proposal 2-1 
For the pattern of SIP of R-TAS, only following 2 alternatives are considered for further down-selection to one alternative:
Alt 1-2: ON-OFF with a ratio of 1:3 and with following total SIP duration to be further down-selected [during RAN1#120bis]:
Option 1: 0.5 OFDM symbol duration
Option 2: 1 OFDM symbol duration
Alt 2-4: ON-OFF-ON-OFF with a ratio of 1:1:1:3 and with following total SIP duration to be further down-selected [during RAN1#120bis]:
Option 1: 0.5 OFDM symbol duration
Option 2: 1 OFDM symbol duration





CAP of R-TAS


[Closed] 1st Discussion Round 
For the CAP of R-TAS, following aspects have been discussed in the contributions following the agreements and discussion from RAN1#120:
Preferred option in terms of variable or fixed duration for CAP design
Preferred exact CAP pattern including M values/duration and CP insertion related aspects

Table 3-1: Summary of views between option 1 and option 2 for CAP




Table 3-2: Summary of views on exact CAP pattern


FL observations
Overall, 30 companies provided their views on CAP of R-TAS
26 companies provided their preference relation to option 1 and option 2 from the TR and 17 companies indicated their preference for option 1, while 8 companies indicated their preference for option 2. 1 company indicated the preference to support and combine both options. Based on the agreement from RAN1#120bis, one option needs to be down-selected in this meeting. From the proponents of option 1, 3 companies provided evaluations and showed that option 1 is able to achieve target FDR, as long as there is sufficient gap between different M values. On the other hand, from the proponents of option 2, 2 companies provided evaluations and showed that at least for higher M values, option 2 is needed. 
Among specific CAP patterns, 5 alternatives are provided. Among the 5 alternatives, majority shows that either 3 chip with alternate ON/OFF or 4 chips with alternate ON/OFF can work well for the M values. Alt 3, 4, 5 with more chips have been proposed only by 1-2 companies each and don’t offer any significant gain compared to Alt 1 and Alt 2 and rather increase the overhead, especially for lower M values
Based on above, proposal 3-1 and proposal 3-2 are provided.



(Updated) Proposal 3-1
For CAP of R-TAS, option 1 for CAP of R-TAS from TR 38.769 is adopted with following details:
CAP duration becomes shorter with increasing value of M
FFS: exact relation between duration of CAP and M values



Proposal 3-2 
For the pattern of CAP of R-TAS, only following 2 alternatives are considered for further down-selection to one alternative:
Alt 1: 3 chips with ON-OFF-ON pattern 
Alt 2: 4 chips with ON-OFF-ON-OFF pattern
Based on the minimum value of M that will be agreed in agenda 9.4.2, corresponding maximum duration of CAP will be selected from the candidate set of {1.5 OFDM symbols, 2 OFDM symbols, 3 OFDM symbols}


[Closed] 2nd Discussion Round 
[Closed] Proposal 3-3
For CAP of R-TAS, following is adopted:
ON-OFF-ON-OFFON pattern with Option 1 for CAP of R-TAS from TR 38.769 is adopted where the CAP duration becomes proportionally shorter with increasing value of M, i.e. if for , duration is  OFDM symbol long, then for , duration is  OFDM symbol long


Proposal 3-3b
For CAP of R-TAS, following is adopted:
ON-OFF-ON-OFFON pattern with Option 1 for CAP of R-TAS from TR 38.769 is adopted where the CAP duration becomes proportionally shorter with increasing value of M, i.e. if for , duration is  OFDM symbol long, then for , duration is  OFDM symbol long
Note: Duration without CP insertion is considered above, with CP insertion, the total duration may not be exactly proportional
Only following two alternatives for CAP pattern  are considered for further down-selection to one alternative:
Alt 1: ON-OFF-ON-OFF
Alt 2: ON-OFF-ON





R2D Postamble



[Closed] 1st Discussion Round 
FL observations
24 companies provided their views in terms of different options for indicating the end of PRDCH transmission
Overall, 11 companies prefer indicating the end of PRDCH transmission based on some unique pattern that violate Manchester coding rule:
7 companies prefer to explicitly specify postamble design
4 companies prefer to not specify postamble design and leave it up to the reader implementation to violate the Manchester coding rule
8 companies prefer indication the end of PRDCH transmission based on R2D control information. Among the 8 companies there are still different views on whether R2D control information is L1 or not
3 companies consider supporting combination of the above 2 options
From FL perspective, all the options work and doesn’t necessarily have any significant performance difference. However, for R2D control indication based method, it is still not clear whether L1 R2D control information is used or not. Based on the proponents, the views are still diverging. If L1 R2D control information is used, then it needs further discussion on separate or joint CRC attachment. On the other hand, for Manchester coding rule based option, in principle companies have common understanding that violation, for example could be based on transmission of three high or three low voltages, i.e. ON-ON-ON or OFF-OFF-OFF. But the main divergence is whether it needs to be specified or not.


Based on above, proposal 4-1 is provided to take the majority view of indicating end of PRDCH transmission based on violation of Manchester coding rule and also with the consideration that with this option, no further discussion is needed on details related to R2D control information and CRC. For method based on violation of Manchester coding rule, furthermore, it can be discussed whether/what postamble needs to be specified or not. If companies cannot converge whether/what postamble needs to be specified or not, then implementation based solution can be adopted.


(Updated) Proposal 4-1 
Take either Alt 1 or Al 2:
Alt 1: For indicating the end of PRDCH transmission, Manchester coding rule violation corresponding to M value is adopted
FFS: whether the Manchester coding rule violation is based on specified postamble pattern or based on device detecting such violation based on reader’s implementation
Note: If companies cannot converge on whether/what postamble is specified or not, then device detecting such violation based on reader’s implementation will be adopted
Note: If no postamble is specified, then from timing perspective, end of last chip of PRDCH is the end of R2D transmission
Alt 2: There is no consensus to specify R2D postamble



[Closed] 2nd Discussion Round 
[Closed] Proposal 4-1b 
At the end of PRDCH, Manchester coding rule violation corresponding to M value is applied by reader’s implementation that can be used by the device to determine the end of PRDCH transmission
Note: From timing perspective, end of last chip of PRDCH is the end of R2D transmission



Proposal 4-1c 
At the end of PRDCH, Manchester coding rule violation corresponding to M value is applied by reader’s implementation that can be used by the device to determine the end of PRDCH transmission
Note: From timing perspective, end of last chip of PRDCH is the end of R2D transmission



Proposed Conclusion 4-2 
There is no consensus to specify R2D postamble



[Closed] 3rd Discussion Round 
(High Priority) Proposal 4-1d 
Manchester coding rule violation corresponding to M value for PRDCH is applied by reader’s implementation that can be used by the device to determine the end of PRDCH transmission
Note: From timing perspective, end of last chip of PRDCH is the end of R2D transmission
There is no consensus to specify R2D postamble








D2R X-ambles (including preamble & midamble)


[Closed] 1st Discussion Round 
Base Sequence Types: First aspect discussed by companies is related to the preference between the two sequence types including the length that were agreed in RAN1#120bis to be further down-selected and corresponding details including sequence length and generation. In Table 5-1 below, the companies’ views are summarized and also including performance analysis, if provided

Table 5-1: Summary of views between M-sequence and Golay Sequence types for D2R X-ambles



Multiple Preamble lengths: Second aspect discussed by companies is related to the number of preamble types in terms of different lengths. In Table 5-2 below, the companies’ views are summarized and also including performance analysis, if provided

Table 5-2: Summary of views on D2R preamble lengths including long and short preamble



# of Preamble Sequences (with same base sequence): Third aspect discussed by companies is related to the number of sequences of same lengths for preamble/midamble. In Table 5-3 below, the companies’ views are summarized and also including performance analysis, if provided

Table 5-3: Summary of views on number of preamble sequences



1- or 2- part Preamble: Fourth aspect is related to 1-part or 2-part preamble design. In Table 5-4 below, the companies’ views are summarized and also including performance analysis, if provided

Table 5-4: Summary of views on number of parts of preamble



Midamble Configuration: Fifth aspect is related to presence/location of D2R x-ambles including whether pre-defined rule and/or explicit signaling by reader is supported. In Table 5-5 below, the companies’ views are summarized and also including performance analysis, if provided

Table 5-5: Summary of views on midamble configuration





Midamble Sequence: Fifth aspect is related to whether midamble sequence is same as preamble or not. In Table 5-6 below, the companies’ views are summarized and also including performance analysis, if provided

Table 5-6: Summary of views on midamble sequence



FL observations
In terms of the base sequence type, all the companies that provided simulation results showed that the performance is similar with both Golay sequence and M sequence. And in terms of preference, almost there is equal number of companies supporting both options. For M sequence, it was pointed out by few companies that the sequence generation is perhaps simpler compared to Golay sequence. Therefore, FL proposal is to adopt M sequence for D2R x-ambles. 

In terms of preamble types, at least 11 companies prefer to support two preamble formats including short preamble format and long preamble format. Multiple companies pointed out the benefit of supporting the two formats and also, one company demonstrated the benefit of using two lengths. In terms of sequence length, 32(-1) bits is proposed by majority of company. For the shorter format, it was shown that 8(-1) bits have reasonable performance difference, while 16-(1) bits have almost similar performance. Therefore, FL proposes to support both short and long preamble format, with length 8(-1) bits and 32(-1) bits, respectively. Also, all but 2 companies considered same sequence for both midamble and preamble as both serve similar functionality and this would require less specification effort. Therefore, same sequence for both is also proposed by FL. In terms of 1-part or 2-part format for preamble, 4 companies prefer to have 2-part for coarse and fine synchronization, respectively. On the other hand, 2 companies explicitly proposed not to support 2-part preamble. One company showed in evaluations that as long as preamble and midamble are designed optimally in terms of length and sequence, there is not benefit from 2-part preamble. In order to get views from more companies, FL will ask companies to provide inputs on their preference

Regarding the configuration of midambles, ~13 companies prefer explicit indication by the network, while 8 companies prefer pre-defined rule to determine number of midambles and location. In terms of pre-defined rule, majority consider supporting TBS based midamble determination. Furthermore, there are 7 companies prefer a combination of pre-defined rule and explicit indication by network. From FL perspective, all solutions work, but it seems there is a quite good number of companies that prefer additional control to reader in terms of at least the presence and number of midamble due to potentially dynamic change in channel conditions and/or reader-specific implementation. Therefore, FL proposal would be to consider the combined option where the number of midambles can be explicitly indicated by the reader to the device via R2D control information and the position of midambles can be at least pre-defined. 


Based on above summary of views from companies on D2R x-ambles, following proposals are provided




(Updated) Proposal 5-1 
For D2R, 1-part preamble and 1-part midamble, base sequence is generated from M-sequence, where the length of the sequence is 
Value(s) of n = 5, 3
Long preamble is generated based on n = 5
Short preamble is generated based on following n values to be down-selected to single value
Alt 1: n = 3
Alt 2: n = 4
FFS: signaling/application of the two preambles
Same base sequence is applied for both preamble and midamble
Preamble immediately precedes the PDRCH without any gap


[Closed] Question 5-1 
Do you support 1-part or 2-part preamble, where for 2-part preamble, 1st part is based on 0s on 1s without Manchester coding and 2nd part is based on the base sequence (as proposed in proposal 5-1)?


(Updated) Proposal 5-2
For D2R midamble, following two alternatives are considered for further down-selection in RAN1#120bis for the determining the presence and location of midamble(s) at the device:
Alt 1: Reader explicitly indicates the number of midambles to be applied by the device for PDRCH transmission via R2D control information
FFS: details of signaling
Based on the indicated number of midambles, a pre-defined rule is applied at the device to determine the exact location of each of the midamble(s) 
FFS: details of pre-defined rule for determining the exact location of midamble(s) 
Alt 2: Reader explicitly indicates the interval between consecutive x-ambles
FFS: details of signaling
Note: These alternatives don’t preclude support of no midamble




[Closed] 2nd Discussion Round 
[Closed] Proposal 5-1b 
For D2R, 1-part preamble and 1-part midamble, base sequence is generated from M-sequence, where the length of the sequence is 
Value(s) of n
Long preamble is generated based on n = 5
Short preamble is generated based on following n values to be down-selected to single value
Alt 1: n = 3
Alt 2: n = 4
FFS: signaling/application of the two preambles
Same base sequence is applied for both preamble and midamble
Preamble immediately precedes the PDRCH without any gap



Proposal 5-1c 
For D2R preamble/midamble, base sequence is generated from m-sequence, where the length of the sequence is 
Value(s) of n
Long preamble/midamble is generated based on n = 5
Short preamble/midamble is generated based on following n values to be down-selected to single value
Alt 1: n = 3
Alt 2: n = 4
Only 1-part preamble/midamble are supported for D2R
Preamble immediately precedes the PDRCH without any gap
If both long and short preamble/midamble are supported, then reader separately indicates 
For preamble, whether long preamble or short preamble is applied
For midamble, whether long midamble or short midamble is applied
For the same length of preamble and midamble, same sequence is supported


(Updated) Proposal 5-2 
For indicating the interval between consecutive midambles, and between the preamble and the first midamble, via R2D control information, following is adopted:
Unit of interval
Alt 1: Number of “pairs of chips”
Alt 2: Number of bits
FFS: the candidate values in terms of the unit of interval


R-TAS (SIP+CAP)
[Closed] Question 6-1
Based on the agreed SIP pattern and CAP pattern alternatives, could companies indicate their preference for the combined design of SIP + CAP, i.e. R-TAS options below? 
R-TAS Option 1: SIP Alt-1-1, Option 1 + CAP Alt-1
R-TAS Option 2: SIP Alt-1-1, Option 1 + CAP Alt-2
R-TAS Option 3: SIP Alt-1-1, Option 2 + CAP Alt-1
R-TAS Option 4: SIP Alt-1-1, Option 2 + CAP Alt-2
R-TAS Option 5: SIP Alt-2-4, Option 1 + CAP Alt-1
R-TAS Option 6: SIP Alt-2-4, Option 1 + CAP Alt-2
R-TAS Option 7: SIP Alt-2-4, Option 2 + CAP Alt-1
R-TAS Option 8: SIP Alt-2-4, Option 2 + CAP Alt-2



(High Priority) Proposal 6-1v1 
For R-TAS, SIP duration of 1 OFDM symbol is adopted with CAP pattern ON-OFF-ON-OFF for all values of M corresponding to PRDCH 

(High Priority) Proposal 6-1v2 
For R-TAS, SIP duration of 0.5 OFDM symbol is adopted with CAP pattern ON-OFF-ON for all values of M corresponding to PRDCH 



Proposals for offline sessions
1st offline session (Monday, April 7, 2025)
(Updated) Proposal 5-2
For D2R midamble, following two alternatives are considered for further down-selection in RAN1#120bis for the determining the presence and location of midamble(s) at the device:
Alt 1: Reader explicitly indicates the number of midambles to be applied by the device for PDRCH transmission via R2D control information
FFS: details of exact signaling to be discussed under agenda 9.4.4
Based on the indicated number of midambles, pre-defined rule(s) are applied at the device to determine the exact location of each of the midamble(s) 
FFS: details of pre-defined rule for determining the exact location of midamble(s)
Alt 2: Reader explicitly indicates the interval between consecutive x-ambles
FFS: details of exact signaling
Note: These alternatives don’t preclude support of no midamble

(Updated) Proposal 3-1
For CAP of R-TAS, option 1 for CAP of R-TAS from TR 38.769 is adopted with following details:
CAP duration becomes shorter with increasing value of M
FFS: Exact relation between duration of CAP and M values

(Updated) Proposal 2-1 
For the pattern of SIP of R-TAS, only following 2 alternatives are considered for further down-selection to one alternative:
Alt 1-2: ON-OFF with a ratio of 1:3 and with following total SIP duration to be further down-selected during RAN1#120bis:
Option 1: 0.5 OFDM symbol duration
Option 2: 1 OFDM symbol duration
Alt 2-4: ON-OFF-ON-OFF with a ratio of 1:1:1:3 and with following total SIP duration to be further down-selected during RAN1#120bis:
Option 1: 0.5 OFDM symbol duration
Option 2: 1 OFDM symbol duration

Proposal 4-1 
For indicating the end of PRDCH transmission, Manchester coding rule violation corresponding to M value is adopted
FFS: whether the Manchester coding rule violation is based on specified postamble pattern or based on device detecting such violation based on reader’s implementation
Note: If companies cannot converge on whether/what postamble is specified or not, then device detecting such violation based on reader’s implementation will be adopted
Note: If no postamble is specified, then from timing perspective, end of last chip of PRDCH is the end of R2D transmission


(Updated) Proposal 5-1 
For D2R, 1-part preamble and 1-part midamble, base sequence is generated from M-sequence, where the length of the sequence is 
Value(s) of n = 5, 3
Long preamble is generated based on n = 5
Short preamble is generated based on following n values to be down-selected to single value
Alt 1: n = 3
Alt 2: n = 4
FFS: signaling/application of the two preambles
Same base sequence is applied for both preamble and midamble
Preamble immediately precedes the PDRCH without any gap

Proposal 3-2 
For the pattern of CAP of R-TAS, only following 2 alternatives are considered for further down-selection to one alternative:
Alt 1: 3 chips with ON-OFF-ON pattern 
Alt 2: 4 chips with ON-OFF-ON-OFF pattern
Based on the minimum value of M that will be agreed in agenda 9.4.1, corresponding maximum duration of CAP will be selected from the candidate set of {1.5 OFDM symbols, 2 OFDM symbols, 3 OFDM symbols}


2nd offline session (Tuesday, April 8, 2025)
Proposal 5-1b 
For D2R, 1-part preamble and 1-part midamble, base sequence is generated from m-sequence, where the length of the sequence is 
Value(s) of n
Long preamble/midamble is generated based on n = 5
Short preamble/midamble is generated based on following n values to be down-selected to single value
Alt 1: n = 3
Alt 2: n = 4
If both long and short preamble/midamble are supported, reader explicitly indicates it
Same base sequence is applied for both preamble and midamble
Preamble immediately precedes the PDRCH without any gap



Proposal 3-3a
For CAP of R-TAS, following is adopted:
ON-OFF-ON-OFFON pattern with Option 1 for CAP of R-TAS from TR 38.769 is adopted where the CAP duration becomes proportionally shorter with increasing value of M, i.e. if for , duration is  OFDM symbol long, then for , duration is  OFDM symbol long
Note: Duration without CP insertion is considered above, with CP insertion the total duration may not be exactly proportional

Proposal 4-1c 
At the end of PRDCH, Manchester coding rule violation corresponding to M value is applied by reader’s implementation that can be used by the device to determine the end of PRDCH transmission
Note: From timing perspective, end of last chip of PRDCH is the end of R2D transmission


Proposed Conclusion 4-2 
There is no consensus to specify R2D postamble


3rd offline session (Wednesday, April 9, 2025)
Proposal 5-1c 
For D2R preamble/midamble, base sequence is generated from m-sequence, where the length of the sequence is 
Value(s) of n
Long preamble/midamble is generated based on n = 5
Short preamble/midamble is generated based on following n values to be down-selected to single value
Alt 1: n = 3
Alt 2: n = 4
Only 1-part preamble/midamble are supported for D2R
Preamble immediately precedes the PDRCH without any gap
Reader explicitly indicates the following cases to the device
Case of indicating short preamble and short midamble is supported
Case of indicating long preamble and long midamble is supported
[Case of indicating long preamble and short midamble is supported]
[For the same length of preamble and midamble, same sequence is supported]


Proposal 3-3b
For CAP of R-TAS, following is adopted:
ON-OFF-ON-OFFON pattern with Option 1 for CAP of R-TAS from TR 38.769 is adopted where the CAP duration becomes proportionally shorter with increasing value of M, i.e. if for , duration is  OFDM symbol long, then for , duration is  OFDM symbol long
Note: Duration without CP insertion is considered above, with CP insertion, the total duration may not be exactly proportional
Only following two alternatives for CAP pattern  are considered for further down-selection to one alternative:
Alt 1: ON-OFF-ON-OFF
Alt 2: ON-OFF-ON

(Updated) Proposal 5-2 
For indicating the interval between consecutive midambles, and between the preamble and the first midamble, via R2D control information, following is adopted:
Unit of interval
Alt 2: Number of bits after FEC (if FEC is applied) and repetition
FFS: the candidate values in terms of the unit of interval

Proposal 4-1c 
At the end of PRDCH, Manchester coding rule violation corresponding to M value is applied by reader’s implementation that can be used by the device to determine the end of PRDCH transmission
Note: From timing perspective, end of last chip of PRDCH is the end of R2D transmission

Proposed Conclusion 4-2 
There is no consensus to specify R2D postamble




4th offline session (Thursday, April 10, 2025)
(High Priority) Proposal 6-1v1 
For R-TAS, SIP duration of 1 OFDM symbol is adopted with CAP pattern ON-OFF-ON-OFF for all values of M corresponding to PRDCH 
Supporting Companies: Samsung, Vivo, Qualcomm, NTT Docomo, Oppo, IDC, Lenovo, China Telecom, NEC, Tejas, Honor, LGE, CEWit, TCL, Spreadtrum, Futurewei, Ericsson, Panasonic, Fujitsu

(High Priority) Proposal 6-1v2 
For R-TAS, SIP duration of 0.5 OFDM symbol is adopted with CAP pattern ON-OFF-ON for all values of M corresponding to PRDCH 
Supporting Companies: Xiaomi, CMCC

(High Priority) Proposal 4-1d 
Manchester coding rule violation corresponding to M value for PRDCH is applied by reader’s implementation that can be used by the device to determine the end of PRDCH transmission
Note: From timing perspective, end of last chip of PRDCH is the end of R2D transmission
There is no consensus to specify R2D postamble


Proposals for online session
1st online session (Tuesday, April 8, 2025)
Proposal 5-2a
For D2R midamble, following two alternatives are considered for further down-selection in RAN1#120bis for determining the presence and location of midamble(s) at the device:
Alt 1: Reader explicitly indicates the number of midambles to be applied by the device for PDRCH transmission via R2D control information
FFS: details of signaling
Based on the indicated number of midambles, pre-defined rule(s) are applied at the device to determine the exact location of each of the midamble(s) 
FFS: details of pre-defined rule for determining the exact location of midamble(s) 
Alt 2: Reader explicitly indicates the interval between consecutive x-ambles
FFS: details of signaling and any pre-define rule(s), if needed
Note: These alternatives don’t preclude the support of no midamble

Proposal 2-1a
For the pattern of SIP of R-TAS, only following 2 alternatives are considered for further down-selection to one alternative:
Alt 1-2: ON-OFF with a ratio of 1:3 and with following total SIP duration to be further down-selected [during RAN1#120bis]:
Option 1: 0.5 OFDM symbol duration
Note: In this option, the chip duration corresponds to M = 8, but this doesn’t imply support of such reference M value for SIP
Option 2: 1 OFDM symbol duration
Note: In this option, the chip duration corresponds to M = 4, but this doesn’t imply support of such reference M value for SIP
Alt 2-4: ON-OFF-ON-OFF with a ratio of 1:1:1:3 and with following total SIP duration to be further down-selected [during RAN1#120bis]:
Option 1: 0.5 OFDM symbol duration
Note: In this option, the chip duration corresponds to M = 12, but this doesn’t imply support of such reference M value for SIP
Option 2: 1 OFDM symbol duration
Note: In this option, the chip duration corresponds to M = 6, but this doesn’t imply support of such reference M value for SIP

Proposal 3-1a
For CAP of R-TAS, option 1 for CAP of R-TAS from TR 38.769 is adopted with following details:
CAP duration becomes shorter with increasing value of M
FFS: exact relation between duration of CAP and M values including {2,4,6,24}

Proposal 3-2a 
For the pattern of CAP of R-TAS, only following 2 alternatives are considered [for further down-selection to one alternative]:
Alt 1: 3 chips with ON-OFF-ON pattern 
For lowest M  = 2, maximum CAP duration will be 1.5 OFDM symbols with this alternative
Alt 2: 4 chips with ON-OFF-ON-OFF pattern
For lowest M  = 2, maximum CAP duration will be 2 OFDM symbols with this alternative
Based on the minimum value of M that will be agreed in agenda 9.4.2, corresponding maximum duration of CAP will be selected from the candidate set of {1.5 OFDM symbols, 2 OFDM symbols, 3 OFDM symbols}


Proposal 5-1a 
For D2R, 1-part preamble and 1-part midamble, base sequence is generated from M-sequence, where the length of the sequence is 
Value(s) of n = 5, 3
Long preamble is generated based on n = 5
Short preamble is generated based on following n values to be down-selected to single value
Alt 1: n = 3
Alt 2: n = 4
FFS: signaling/application of the two preambles
Same base sequence is applied for both preamble and midamble
Preamble immediately precedes the PDRCH without any gap



Proposal 4-1a 
Take either Alt 1 or Al 2:

Alt 1: For indicating the end of PRDCH transmission, Manchester coding rule violation corresponding to M value is adopted
FFS: whether the Manchester coding rule violation is based on specified postamble pattern or based on device detecting such violation based on reader’s implementation
Note: If companies cannot converge on whether/what postamble is specified or not, then device detecting such violation based on reader’s implementation will be adopted
Note: If no postamble is specified, then from timing perspective, end of last chip of PRDCH is the end of R2D transmission
Alt 2: There is no consensus to specify R2D postamble

2nd online session (Wednesday, April 9, 2025)
Proposal 5-1d 
For D2R preamble/midamble, base sequence is generated from m-sequence, where the length of the sequence is 
Value(s) of n
Long preamble/midamble is generated based on n = 5
Short preamble/midamble is generated based on following n values to be down-selected to single value
Alt 1: n = 3
Alt 2: n = 4
Only 1-part preamble/midamble are supported for D2R
Preamble immediately precedes the PDRCH without any gap
If both long and short preamble and midamble (if present) are supported, then following cases are supported and reader explicitly  indicates one of the following cases for PDRCH:
Short preamble and short midamble 
Long preamble and long midamble 
[Long preamble and short midamble]
[For the same length of preamble and midamble, same sequence is supported]

Proposal 3-3c
For CAP of R-TAS, following is adopted:
Option 1 for CAP of R-TAS from TR 38.769 is adopted where the CAP duration becomes proportionally shorter with increasing value of M, i.e. if for , duration is  OFDM symbol long, then for , duration is  OFDM symbol long
Note: Duration without CP insertion is considered above, with CP insertion, the total duration may not be exactly proportional
Only following two alternatives for CAP pattern  are considered for further down-selection to one alternative:
Alt 1: ON-OFF-ON-OFF
Alt 2: ON-OFF-ON

Proposal 5-2a 
For indicating the interval between consecutive midambles, and between the preamble and the first midamble, via R2D control information, following is adopted:
Unit of interval is number of bits after FEC (if FEC is applied) and repetition
FFS: the candidate values in terms of the unit of interval
Proposal 4-1c 
At the end of PRDCH, Manchester coding rule violation corresponding to M value is applied by reader’s implementation that can be used by the device to determine the end of PRDCH transmission
Note: From timing perspective, end of last chip of PRDCH is the end of R2D transmission

Proposed Conclusion 4-2 
There is no consensus to specify R2D postamble


3rd online session (Thursday, April 10, 2025)
Either Take v1 or v2 for Proposal 6-1
(High Priority) Proposal 6-1v1 
For R-TAS, SIP duration of 1 OFDM symbol is adopted with CAP pattern ON-OFF-ON-OFF for all values of M corresponding to PRDCH 
Supporting Companies (19 companies): Samsung, Vivo, Qualcomm, NTT Docomo, Oppo, IDC, Lenovo, China Telecom, NEC, Tejas, Honor, LGE, CEWit, TCL, Spreadtrum, Futurewei, Ericsson, Panasonic, Fujitsu

(High Priority) Proposal 6-1v2 
For R-TAS, SIP duration of 0.5 OFDM symbol is adopted with CAP pattern ON-OFF-ON for all values of M corresponding to PRDCH 
Supporting Companies (2 companies): Xiaomi, CMCC

(High Priority) Proposal 4-1d 
Manchester coding rule violation corresponding to M value for PRDCH is applied by reader’s implementation that can be used by the device to determine the end of PRDCH transmission
Note: From timing perspective, end of last chip of PRDCH is the end of R2D transmission
There is no consensus to specify R2D postamble









Contributions in RAN1#120bis







Appendix
Revised WID (RP-243326): RAN1 Scope & Objectives 
General Scope
The definitions provided in TR 38.848, TR 38.769, and decisions, etc. made during the Rel-19 SI in RAN WGs are taken into this WI, and the following is the exclusive general scope:
The overall objective shall be to standardize the following Ambient IoT device:
Device 1: ~1 µW peak power consumption, has energy storage, RF envelope detector receiver, initial sampling frequency offset (SFO) up to 10X ppm, neither R2D nor D2R amplification in the device. The device’s D2R transmission is backscattered on a carrier wave provided externally.
Deployment scenario 1 with Topology 1, according to D1T1-B. 
FR1 licensed spectrum in FDD, with R2D in DL spectrum and D2R and CW in UL spectrum.
Spectrum deployment in-band to NR and standalone, with A-IoT BS located indoor.
Traffic types DO-DTT, DT, for rUC1 (indoor inventory) and rUC4 (indoor command). 
Carrier wave transmission for waveform 1 only, without hopping, per the following cases in TR 38.769:
Case 1-4 for D1T1-B
Proximity determination via Solution 1 in TR 38.769 only.
Device (un)availability via Direction 1 in TR 38.769 only.

WGs begin their discussions from the decisions already made in TR 38.769, with the following refinements for the scope: 

The following objectives are set, within the General Scope:
RAN1 scope:
PRDCH and PDRCH, which are the only physical channels in R2D and D2R, respectively.
R2D and D2R signal(s)
Multiplexing/multiple access in R2D is by only TDMA, and in D2R is by only TDMA and FDMA.
R2D supports only OOK-4 modulation, one solution for CP handling. D2R backscattering supports only OOK and BPSK modulations.
R2D transmission supports only the Manchester line code in TR 38.769
D2R transmission supports:
Either the Manchester line code in TR 38.769 or no line code (one to be down-selected); and
A corresponding small frequency shift method according to the options in TR 38.769.
R2D does not support FEC. D2R supports only convolutional code with generator polynomials as per TS 36.212. Applying or not applying the FEC to D2R is specified by ensuring it is under the reader control and applies to all devices targeted by the reader.
PRDCH and PDRCH both support transmission without CRC, and with CRC as per the generator polynomials in TS 38.212 for 6-bit CRC and 16-bit CRC. Cases to use which length of CRC, or no CRC, to be decided in RAN1.
D2R supports physical layer repetition transmission. R2D does not support physical-layer repetition transmission. 
RAN2 scope:
Specify the necessary functions and procedures for an Ambient IoT compact protocol stack and lightweight signalling procedure to enable DO-DTT and DT data transmission:
A-IoT Paging, including subsequent paging for the same service. Support the options that a paging message contains one identifier, and that a paging message contains no identifier. 
Note: RAN2 aims to design a paging message format such that multiple identifiers can be contained in one paging message, for forward compatibility purposes.
A-IoT Random access, including re-access for failure handling. Contention-based and contention-free cases are supported. For the contention-based random access, only Solution 1 (3-step only) is included.
A-IoT data transmission, including data (re-)transmission for failure handling. Segmentation is supported at least in D2R.
Only MAC layer is included
RAN3 scope: 
Specify necessary architectural aspects, and signaling and procedures between A-IoT RAN and A-IoT CN to support the A-IoT functions, assuming an architecture of aggregated gNB, including:
Inventory and command operations
Device location reporting at reader ID granularity
Note: The above A-IoT functions are supported over the existing NG interface, based on architecture(s) defined by RAN3/SA2.
RAN4 scope:
Specify RF requirements for Ambient-IoT BS, device 1, and CW
RF requirements for Type 1-C Ambient-IoT BS
RF requirements for device 1
RF requirements for CW
Specify RRM core requirements for device 1, if necessary
Study and develop OTA test methodology for A-IoT device 1
Consider test methods specified in TR 38.870 as starting point. Take test system reuse, test system complexity and test time into account, when developing test methods suitable for Ambient IoT.
Develop the preliminary Measurement Uncertainty (MU) assessment for the test system
Use band n8 as an example band

Note 1: Coordination with SA2 and SA3 is expected. Updates to the WID objectives should be considered if needed.
Note 2: This WI shall target for an IoT segment well below the existing 3GPP IoT technologies, e.g. NB-IoT, eMTC, RedCap, etc. The WI shall not aim to replace existing 3GPP LPWA technologies.

SI Phase: RAN1 Agreements (relevant for R2D/D2R signals including timing acquisition and synchronization
RAN1#116 (Athens, Greece, February 26th – March 1st, 2024)
Agreement
At least the following time domain frame structure is studied for A-IoT R2D and D2R transmission.
For R2D transmission,
A R2D timing acquisition signal (e.g. R2D preamble) is included at least for timing acquisition and for indicating the start of the R2D transmission in time domain.
For D2R transmission,
A D2R timing acquisition signal (e.g. D2R preamble) is included at least for timing acquisition and for indicating the start of the D2R transmission in time domain.
FFS other necessary component(s), e.g. midamble, postamble, periodic sync signal, control fields, guard period


RAN1#116bis (Changsha, Hunan Province, China, April 15th – April 19th, 2024)

Agreement
To determine or derive the end of PRDCH transmission, study at least following options:  
Option 1: R2D postamble immediately follows the PRDCH to indicate the end of the PRDCH.       
Option 2: Based on R2D control information.

Agreement
For the reader to acquire the end of PDRCH transmission, study at least following options:  
Option 1: D2R postamble immediately follows the PDRCH
Option 2: Based on control information

Agreement
For D2R transmission, study the necessity of midamble at least for the purpose of performing timing/frequency tracking or channel estimation or interference estimation, considering at least the following: 
Modulation and Coding schemes, e.g., data modulation, line/channel coding 
Receiving methods, e.g., coherent or non-coherent
D2R transmission length/packet size
Midamble overhead
Timing/frequency accuracy
Phase accuracy

Agreement
RAN1 study the R2D transmission without midamble as the baseline if Manchester encoding is used.
FFS the necessity for the R2D transmission with midamble if PIE is used. 

Agreement
For the R2D timing acquisition signal immediately preceding the transmission of a physical channel, study a preamble with at least two parts which includes a start-indicator part and a clock-acquisition part, where the start-indicator part immediately precedes the clock-acquisition part:
Start-indicator part provides the start of the R2D transmission
FFS: Details of start-indicator part
Clock-acquisition part provides at least the chip synchronization of the subsequent physical channel transmission
FFS: Details of clock-acquisition part, e.g. structure, encoding, length, etc. 
FFS: Methods to determine chip duration of the subsequent physical channel transmission 
FFS: Other functionalities
Note: the preamble is considered not to be part of a physical channel
FFS: other part(s) of the preamble, if any 
FFS: whether the above clock acquisition is sufficient for all devices
FFS: how to make the preamble compact

Agreement
For D2R, a preamble preceding each PDRCH transmission is studied as the baseline at least for the D2R timing acquisition signal:
Preamble is not part of PDRCH
FFS: Other functionalities of the preamble

Agreement
Reference signals including at least DMRS, PTRS, CSI-RS/TRS, are not further studied for R2D.

Agreement
Reference signals including DMRS, PTRS, SRS, are not further studied for D2R
Note: This doesn’t preclude the possibility to study preamble, midamble, postamble for different purposes, e.g. channel/interference estimation and/or proximity determination



RAN1#117 (Fukuoka City, Fukuoka, Japan, May 20th – 24th, 2024)

Agreement
For the start-indicator part of the R2D time acquisition signal, study the two options below:
Option 1: ON/OFF pattern i.e. high/low voltage transmission 
Option 2: OFF pattern, i.e. low voltage transmission 

Agreement
For R2D, the clock-acquisition part of the R2D time acquisition signal is used to determine the OOK chip duration
FFS: Pattern design to support determination of chip duration


RAN1#118 (Maastricht, NL, August 19th – 23rd,  2024)

Agreement
For each D2R transmission, no separate part for start-indicator is considered for the preamble preceding the PDRCH.

Agreement
For D2R transmission, preamble preceding the PDRCH is studied also for the potential additional functionalities:
SFO estimation
CFO estimation
Channel estimation
Interference estimation
Note: this does not preclude studying the above functionalities by using a midamble and/or postamble, if supported
FFS: Other functionalities, if any

Agreement
For the start-indicator part of the R2D time acquisition signal, ON/OFF pattern i.e. high/low voltage transmission is applied
FFS: length/pattern of ON/OFF.
FFS: when TD2R_min is applicable, whether/how the start-indicator part is included in TD2R_min or not. To be discussed in 9.4.2.2


RAN1#118bis (Hefei, China, October 14th – 18th,  2024)

Agreement
The start indicator part of the R2D time acquisition signal is not included in TD2R_min.

Agreement
The TR will capture the following options, and companies are encouraged to analyze the tradeoffs among the following D2R amble(s) options:
Option 1: D2R preamble only
Option 2: D2R preamble + X midamble(s), where X 1
Option 3: D2R preamble + postamble
Option 4: D2R preamble + Y midamble(s) + postamble, where Y1
For the above options, companies are encouraged to report at least the following:
Purpose(s) of the preamble, midamble and postamble 
Whether companies assume multiple options can be supported


Agreement
For analysing the trade-offs among the D2R amble(s) options, companies can refer to the Table 3.2.4 in section 3.2.4 of R1-2408993 for information.

Agreement
For the clock-acquisition part of the R2D time acquisition signal, following is captured in the TR 38.769:
Clock-acquisition part is based on OOK without line coding and includes rising/falling edges, including at least two rising or at least two falling edges for the device to determine the OOK chip duration

Agreement
For the start-indicator part of the R2D time acquisition signal, for providing the start of the R2D transmission, following is captured in the TR 38.769:
Following options have been studied for the start-indicator part of the R2D time acquisition signal:
Option 1: ON-OFF transmission is considered based on energy/edge detection, and multiple alternatives have been studied including 
Alt 1: A single ON-OFF transmission, i.e. one high-voltage transmission followed by one low-voltage transmission, where ON and OFF may have same or different durations
Alt 2: A multi-ON-OFF transmission, where different ON and different OFF may have same or different durations and different parts may have same or different duration
Option 2: ON-OFF sequence-based design is considered which consists of a pre-defined sequence for detection of start-indicator part based on digital correlation
For both the options, it is observed that a fixed duration for the start-indicator part can be considered, regardless of the value of M used for PRDCH transmissions. 
Miss-detection ratio (MDR), false-alarm ratio (FAR) and detection complexity have been considered for the design of the R2D start indicator part by following companies
It is observed by 1 source [Huawei] that for an FAR of ~0%, the MDR of less than 1% can be achieved with Alt 2 of option 1 (considering 2 ON-OFF transmissions with different durations) and it is also observed that low-complexity and reduced power consumption can be achieved
1 source [ZTE] evaluated Alt 1 of option 1 (considering same duration for ON and OFF) and Alt 2 of option 1 (considering multiple ON-OFF transmissions with same duration) and observed that for an FAR of ~0%, the MDR of less than 1% can be achieved and Alt 1 of option 1 performs better than Alt 2 of option 1. 
1 source [CATT] observed with ON-OFF pattern, that for an FAR of ~0%, the MDR of less than 1% can be achieved with a duration of at least 1 OFDM symbol
1 source [Qualcomm] compares the performance between option 1 and option 2. It shows almost similar coverage range (SNR requirement) for target MDR of 1%. For MDR of 10%, it shows that sequence-based design provides better performance, and it is observed that during the available time, it is feasible for all devices to detect the start-indicator sequence within the power budget. It is further observed that the FAR with sequence-based design can be improved in case of interference scenarios when compared with pattern-based design. 
For both the options, it may be beneficial that the start-indicator part is distinguishable at least from other parts of the R2D transmissions

Agreement
For the clock-acquisition part of the R2D time acquisition signal for OOK chip duration determination, following options are studied:
Option 1: Duration of the clock-acquisition part is variable for different M values, i.e. the duration becomes shorter with increasing value of M
Option 2: Duration of the clock-acquisition part is constant for different M values based on repetition, i.e. repetition factor is increased with increasing value of M to keep the duration constant
FFS: Whether/what restriction on M values for the clock-acquisition part
Note: Other functionalities of clock-acquisition part is a separate discussion

Agreement
For the D2R preamble, binary signal is considered.


RAN1#119 (Orlando, US, Nov 18th – 22nd, 2024)
Agreement
Capture following observations in the TR 38.769, where CFO is assumed to be zero or negligible.
For coherent detection of PDRCH with a payload of 16 bits or 20 bits with 6-bit or 16-bit CRC, using 1/2 Manchester coding and 1/3 or 1/2 convolutional code:
Source [6, ZTE] observed that with up to 10% SFO, ~1kbps data rate, when the same amble(s) overhead is maintained, Option 3 provides comparable performance results to Option 1.
Source [7, Samsung] observed that with up to 10% SFO, ~5kbps data rate, for device 1 and with up to 1% SFO for device 2, the decoding performance with/without midamble are similar
Source [9, vivo] observed that Option 1 is sufficient to achieve 10% and 1% BLER, with no more than 8 SFO hypotheses tested at the reader side.
With up to 10% SFO, ~ 5kbps data rate, the SNR needed to achieve 10% and 1% BLER is similar (~ -2dB and 4 dB) for Option 1, Option 2 of D2R preamble+1midamble and Option 3.
With up to 1% SFO, ~ 5kbps data rate, the SNR needed to achieve 10% and 1% BLER is similar (~ -2.8dB and 3.3dB) for Option 1, Option 2 of D2R preamble+1 midamble and Option 3.
Source [20, OPPO] observed that with perfect SFO estimation, 1kbps data rate and OOK modulation, there is no noticeable performance gain from using midamble(s) and/or postamble for PDRCH transmission, regardless of the payload size.
Source [30, Qualcomm] observed that the required SFO estimation accuracy to achieve 1% and 10% BLER is <5*10^3 ppm
With up to 10% SFO, achieving the required accuracy necessitates more than 20 SFO hypotheses at the reader side for Option 1 and 10 SFO hypotheses are sufficient for Option 3 of D2R preamble + postamble. But for Option 3 reader has to store the received samples and wait for the postamble that is after the end of PDRCH for any of SFO/channel/interference estimation, demodulation, decoding, etc. 
With up to 1% SFO, 4 SFO hypotheses are sufficient for Option 1 to achieve the required accuracy.

For coherent detection of PDRCH with a payload of 96bits with 16-bit CRC (or 6-bit CRC [14, Xiaomi]), using 1/2 Manchester coding and 1/3 or 1/2 convolutional code,
Sources [3, Huawei], [5, CMCC] and [14, xiaomi] observed that Option 1 cannot achieve 10% BLER.
Sources [6, ZTE], [7, Samsung], [9, vivo], [20, OPPO] and [30, QC] observed that Option 1 can achieve 10% BLER.
Sources [3, Huawei], [5, CMCC], [6, ZTE], [7, Samsung], [9, vivo], [14, xiaomi], [16, China Telecom] observed that adding additional amble improves the performance. 
Source [3, Huawei] observed that with up to 10% SFO, 
Option 2 of D2R preamble+ 1 midamble achieves 10% BLER at SNR around -3dB, but cannot achieve 1% BLER.
Option 3 of D2R preamble+ postamble achieves 10% BLER at SNR around -4dB, and can achieve 1% BLER at SNR around 4dB.
Source [5, CMCC] observed that with up to 10% SFO, Option 3 allows reader to precisely search and detect the SFO with 0.03% residual SFO at -3dB SNR TDL-A channel, achieving 10% BLER -2.44dB SNR for ~1 kbps data rate and -2.17 dB for ~6 kbps data rate. Source [5, CMCC] further observed that when the reader adopts same number of SFO hypothesis based on preamble, with 1% SFO, Option 3 can achieve 10% BLER at -4.27 dB SNR for ~1 kbps and at -4.29 dB SNR for ~6 kbps, which provides 1~2 dB performance gain when compared to 10% SFO. 
Source [14, xiaomi] observed that with up to 10% SFO, Option 2 of D2R preamble + 1 midamble, option 3, and option 4 of D2R preamble + 1 midamble+postamble achieve basically the same performance, the SNR for 10% BLER is 5dB for 1.25 kbps data rate.  
Source [6, ZTE] observed that with up to 10% SFO, ~1kbps data rate, and the same amble(s) overhead, Option 3 can provide 1~2 dB, 5dB performance gain for 10% BLER, 1% BLER, respectively, compared to Option 1. Additionally, Option 3 can provide ~1dB, 2dB performance gain for 10% BLER, 1% BLER, respectively, compared to Option 2 of D2R preamble +1 midamble.
Source [7, Samsung] observed that for ~5kbps data rate, compared to option 1, 
For device 1 with up to 10% SFO, Option 2 of D2R preamble + 1 midamble provides ~0.5 dB SNR gain at 10% BLER with TDL-A channel and ~0.9 dB SNR gain with TDL-D channel.
For device 2 with up to 1% SFO, Option 2 of D2R pramble + 1midamble provides ~1 dB SNR gain at 10% BLER with TDL-A channel and ~1.4 dB SNR gain with TDL-D channel.
Source [9, vivo] observed that, 
With up to 10% SFO, ~5.5kbps data rate, Option 1 achieves 10% and 1% BLER at SNR ~0.7dB and 10dB, respectively; Additionally, maintaining the same amble overhead, Option 2 (D2R preamble + 1 midamble) and Option 3 demonstrate similar performance, achieving 10% and 1% BLER at SNR around -1.7dB and 5.2dB, respectively.
With up to 1% SFO, ~5.5kbps data rate, Option 1 achieves 10% and 1% BLER at SNR ~ -1.3dB and 11dB, respectively. Additionally, with the same amble overhead, the SNR difference between Option 2 (D2R preamble+1midamble) and Option 3 for 10% and 1% BLER is less than 1dB, with SNRs ~ -3.1dB to -2.5dB for 10% BLER and ~3.6dB to 4.5dB for 1% BLER.
Source [16, China Telecom] observed that with up to 10% SFO, ~7.5kbps data rate, there is ~6~7dB performance gap at 10% BLER and ~10.5~11.5dB performance gap at 1% BLER between option 2 of D2R preamble+111 midambles and option 1. Note that Source [16, China Telecom] does not use any convolutional code.
Source [20, OPPO] observed that with perfect SFO estimation, 1kbps data rate and OOK modulation, there is no noticeable performance gain from using midamble(s) and/or postamble for PDRCH transmission, regardless of the payload size.
Source [30, Qualcomm] observed that the required SFO estimation accuracy to achieve 1% and 10% BLER is <10^3 ppm. To achieve the required accuracy,
For Option 1, more than 50 SFO hypotheses at reader side are necessary for device with up to 10% SFO and 6 SFO hypotheses are sufficient at reader side for device with up to 1% SFO. 
For Option 3, 10 SFO hypotheses are sufficient for device with up to 10% SFO, but reader has to store the received samples and wait for the postamble that is after the end of PDRCH for any of SFO/channel/interference estimation, demodulation, decoding, etc.

For coherent detection of PDRCH with a payload of 400bits with 16-bit CRC, using 1/2 Manchester coding and 1/3 or 1/2 convolutional code, 
For option 1 of D2R preamble only, 
Sources [3, Huawei], [5, CMCC], [6, ZTE], [8, Spreadtrum], [9, vivo], [14, xiaomi] observed that with up to 10% SFO, 10% BLER cannot be achieved. 
Source [20, OPPO] observed that with perfect SFO estimation, 1kbps data rate and OOK modulation, there is no noticeable performance gain from using midamble(s) and/or postamble for PDRCH transmission.
For other amble options, 
Source [3, Huawei] observed that
With accurate SFO estimation, Option 2 of D2R preamble + 4 midambles can achieve 10% BLER at SNR ~ 2.7dB but cannot achieve 1% BLER.
With up to 10% SFO, Option 3 cannot achieve 10% BLER.
With up to 10% SFO, Option 4 of D2R preamble+2 midambles+postamble achieves 10% BLER at SNR of ~0.25dB; But it cannot achieve 1% BLER. Option 4 of D2R preamble+3 or 4 midambles+postamble, achieves a 10% BLER at an SNR of around -0.2 dB, and achieves 1% BLER at SNR around 9dB or 8dB, respectively.
Source [5, CMCC] observed that with up to 10% SFO, Option 4 of D2R preamble combined with 1 to 4 midambles + postamble, achieves 10% BLER at SNR of 2.5 dB, 1 dB, 0.8 dB, or 0.5 dB, respectively, for a data rate of around 1 kbps.
Source [6, ZTE] observed that with up to 10% SFO, 
Option 3 can provide ~5.5 dB performance gain compared to option 2 of D2R preamble+1midamble for 10% BLER, with the same amble(s) overhead for ~1kbps data rate.
Option 2 of D2R preamble+1midamble cannot achieve 1% BLER for ~1kbps data rate.
Option 4 of the D2R preamble+1 or 2 midamble(s)+postamble, has similar performance, it can achieves a 10% BLER at SNR of -1dB and achieves a 1% BLER at SNR of 6dB and 5dB respectively for ~1kbps data rate.  
Source [8, Spreadtrum] observed that with up to 10% SFO, 
Option 3 of D2R preamble+ postamble cannot achieve 10% BLER for ~7kpbs.  
Option 4 of D2R preamble + 1 midamble + postamble can achieve 10% BLER and 1% BLER at SNR around -6dB and 0 dB, respectively for ~7kpbs data rate.
Source [9, vivo] observed that 
With up to 10% SFO, maintaining the same amble overhead, both Option 2 of D2R preamble+5 midambles and Option 4 of D2R preamble+4 midambles+1 postamble demonstrate similar performance, achieving 10% BLER at SNR ~0.2dB. While for 1% BLER, the SNR for Option 2 and Option 4 is ~ 9.2dB and 12.8dB, respectively for ~5.5kpbs data rate.   
With up to 1% SFO, maintaining the same amble overhead, both Option 2 of D2R preamble+5 midambles and Option 4 of D2R preamble+4 midambles+1postamble demonstrate similar performance, achieving 10% BLER at SNR around -1.2dB. While for 1% BLER, the SNR for Option 2 and Option 4 is ~7.8dB and 9.1dB, respectively for ~5.5kpbs data rate.   
Source [14, xiaomi] observed that with up to 10% SFO, Option 2 of D2R preamble+3 midambles and Option 4 of D2R preamble+3 midambles+postamble can achieve 10% BLER when the SNR is within the range of [15, 25] dB for 1.25 kbps data rate. 
Source [30, Qualcomm] observed that the required SFO estimation accuracy to achieve 1% and 10% BLER is much smaller than 10^3 ppm. To achieve the required accuracy,
For Option 2 of D2R preamble+X midamble(s) where midamble inserted per every certain number of PDRCH bits (e.g., 192 bits),
For SFO estimation using each amble for the subsequent PDRCH bits (e.g., 192 bits), with up to 10% SFO, more than 50 SFO hypotheses are necessary at the reader side and with up to 1% SFO, 6 SFO hypotheses are sufficient at the reader side.
For SFO estimation based on the time gap between preamble and midamble, with up to 10% SFO, 10 SFO hypotheses are used, but reader has to store the received samples and wait for the midamble to start SFO/channel/interference estimation, demodulation, decoding, etc.
For Option 3 of D2R preamble+postamble, SFO estimation is based on the time gap between preamble and postamble, with up to 10% device SFO, 10 SFO hypotheses are used for reader, but reader has to store the received samples and wait for the postamble that is after the end of PDRCH for any of SFO/channel/interference estimation, demodulation, decoding, etc

For the synchronization and timing tracking of D2R transmission, 
Source [5, CMCC] report that with up to 10% SFO, option 1 is not sufficient for D2R reception since the residual SFO at reader side is larger than 1%. While with option 3, the reader can precisely search and detect the SFO with a residual SFO of 0.03% at -3dB SNR TDL-A channel.
Source [14, xiaomi] report that 
For packet size of 96bits, when the SNR is increased from -4dB to 20dB, the ratio of device residual SFO over 100ppm decreases to 6% for Option 2, 3 and 4, but remains at 95% for Option 1.
For packet size of 400bits, when the SNR is increased from -4dB to 20dB, the ratio of device residual SFO larger than 10ppm decreases to 5% for Option 2, 3, and 4, but is still 99.6% for Option 1.
Sources [9, vivo], [15, CATT] report that SFO estimation based on D2R preamble can achieve accurate estimation without additional ambles (midamble or postamble). 
Source [9, vivo][7 Samsung] observed that for non-coherent detection of PDRCH, the number of SFO hypotheses and the SNR needed for 10% and 1% BLER cannot significantly be reduced for option 2, 3 and 4 compared to the option 1. Moreover, the additional ambles i.e., midamble(s) and/or postamble introduces additional overhead and prevents pipelined processing of the reception.     
Source [15, CATT] observed that 
The coarse estimation of SFO based on the D2R preamble indicates that the SFO estimation error is less than 1% with a probability of 99.3%, and less than 0.1% with a probability of 49.9%.
The fine estimation of SFO based on the D2R preamble shows that the SFO estimation error is less than 1% with a probability of 99.5%, and less than 0.1% with a probability of 90.8%.
Reader/gNB can achieve a probability of not less than 99.5% for SFO estimation error below 1%, and 90.8% for SFO estimation error below 0.1% by receiving D2R preamble signals.
Source [30, Qualcomm] report that for D2R with coherent demodulation at reader, the reader needs to estimate the device clock frequency with the accuracy of 0.5% (5 * 10^3 ppm) or lower for a short message (e.g., 72 bits after CRC/coding) and 0.1% (10^3 ppm) or lower for a long message (e.g., 224 bits after CRC/coding). The source further reports that design of D2R amble(s) (e.g., overhead) and the correspondingly required number of SFO hypothesis for the estimation depend on the sampling clock accuracy that the device uses for D2R. 
Note: in the observations above where coherent detection is used, sources that evaluated option 3 and option 4 assumed that the postamble is used at least for time/frequency tracking and for channel estimation.

Agreement
For the CFO calibration signal, which is required only for device 2b to reduce the frequency offset range and the guard-bandwidth of D2R transmission, the following observations are captured in TR 38.769:
Source [3, Huawei] report that a single-tone RF signal is used as the CFO calibration signal, it is not a part of time acquisition signal and can be transmitted as an optional R2D signal after the PRDCH transmission. 
Sources [2, Ericsson], [19, Panasonic] and [20, OPPO] report that additional synchronization signal is needed. 
[OPPO] state the R2D timing acquisition signal may not be sufficient or may not be usable for CFO calibration since a reference frequency is needed when separate LOs are used for Tx and Rx in device 2b.
Sources [7, Samsung], [9, vivo], [30, Qualcomm], [36, Apple] report that additional synchronization signal is needed if the synchronization for carrier frequency using R2D signal/channel does not provide required functionalities for device 2b.
Source [5, CMCC][31, MTK] report that it may not be possible to achieve enough frequency accuracy (0.01 ppm) even after CFO calibration based on R2D time acquisition signals for coherent detection at reader especially when the D2R data rate is low.

Agreement
For device 2b, a signal for CFO calibration should be provided to synchronize / calibrate the device clock for LO for carrier frequency (Clock purpose #5) to achieve the accuracy after clock sync / calibration at device side captured in Table 5.2.3-1.
Frequency calibration at device 2b is beneficial at least to reduce the guard-bandwidth of D2R transmission.

Agreement
Adopt the updates documented in R1-2410653 for section 6.2 of the TR38.769. 

Agreement
Adopt following update to the TP agreed on Monday

Capture following observations in the TR 38.769, where CFO is assumed to be zero or negligible.
[omit unchanged part]
For coherent detection of PDRCH with a payload of 96bits with 16-bit CRC (or 6-bit CRC [14, Xiaomi]), using 1/2 Manchester coding and 1/3 or 1/2 convolutional code,
Sources [3, Huawei], [5, CMCC] and [14, xiaomi] observed that Option 1 cannot achieve 10% BLER.
Sources [6, ZTE], [7, Samsung], [9, vivo], [20, OPPO] and [30, QC] observed that Option 1 can achieve 10% BLER.
Sources [3, Huawei], [5, CMCC], [6, ZTE], [7, Samsung], [9, vivo], [14, xiaomi], [16, China Telecom] observed that adding additional amble improves the performance. 
Source [3, Huawei] observed that with up to 10% SFO, 
Option 2 of D2R preamble+ 1 midamble achieves 10% BLER at SNR around -3dB, but cannot achieve 1% BLER.
Option 3 of D2R preamble+ postamble achieves 10% BLER at SNR around -4dB, and can achieve 1% BLER at SNR around 4dB.
Source [5, CMCC] observed that with up to 10% SFO, Option 3 allows reader to precisely search and detect the SFO with 0.03% residual SFO at -3dB SNR TDL-A channel, achieving 10% BLER -2.44dB SNR for ~1 kbps data rate and -2.17 dB for ~6 kbps data rate. Source [5, CMCC] further observed that when the reader adopts same number of SFO hypothesis based on preamble, with 1% SFO, Option 3 can achieve 10% BLER at -4.27 dB SNR for ~1 kbps and at -4.29 dB SNR for ~6 kbps, which provides 1~2 dB performance gain when compared to 10% SFO. 
Source [14, xiaomi] observed that with up to 10% SFO, Option 2 of D2R preamble + 1 midamble, option 3, and option 4 of D2R preamble + 1 midamble+postamble achieve basically the same performance, the SNR for 10% BLER is 5dB for 1.25 kbps data rate.  
Source [6, ZTE] observed that with up to 10% SFO, ~1kbps data rate, and the same amble(s) overhead, Option 3 can provide 1~2 dB, 5dB performance gain for 10% BLER, 1% BLER, respectively, compared to Option 1. Additionally, Option 3 can provide ~1dB, 2dB performance gain for 10% BLER, 1% BLER, respectively, compared to Option 2 of D2R preamble +1 midamble.
Source [7, Samsung] observed that for ~5kbps data rate, compared to option 1, 
For device 1 with up to 10% SFO, Option 2 of D2R preamble + 1 midamble provides ~0.5 dB SNR gain at 10% BLER with TDL-A channel and ~0.9 dB SNR gain with TDL-D channel.
For device 2 with up to 1% SFO, Option 2 of D2R pramble + 1midamble provides ~1 dB SNR gain at 10% BLER with TDL-A channel and ~1.4 dB SNR gain with TDL-D channel.
Moreover, the additional ambles i.e., midamble(s) and/or postamble introduces additional overhead and postamble may prevent pipelined processing of the reception. 
Source [9, vivo] observed that, 
With up to 10% SFO, ~5.5kbps data rate, Option 1 achieves 10% and 1% BLER at SNR ~0.7dB and 10dB, respectively; Additionally, maintaining the same amble overhead, Option 2 (D2R preamble + 1 midamble) and Option 3 demonstrate similar performance, achieving 10% and 1% BLER at SNR around -1.7dB and 5.2dB, respectively.
With up to 1% SFO, ~5.5kbps data rate, Option 1 achieves 10% and 1% BLER at SNR ~ -1.3dB and 11dB, respectively. Additionally, with the same amble overhead, the SNR difference between Option 2 (D2R preamble+1midamble) and Option 3 for 10% and 1% BLER is less than 1dB, with SNRs ~ -3.1dB to -2.5dB for 10% BLER and ~3.6dB to 4.5dB for 1% BLER.
Source [16, China Telecom] observed that with up to 10% SFO, ~7.5kbps data rate, there is ~6~7dB performance gap at 10% BLER and ~10.5~11.5dB performance gap at 1% BLER between option 2 of D2R preamble+111 midambles and option 1. Note that Source [16, China Telecom] does not use any convolutional code.

Source [20, OPPO] observed that with perfect SFO estimation, 1kbps data rate and OOK modulation, there is no noticeable performance gain from using midamble(s) and/or postamble for PDRCH transmission, regardless of the payload size.
Source [30, Qualcomm] observed that the required SFO estimation accuracy to achieve 1% and 10% BLER is <10^3 ppm. To achieve the required accuracy,
For Option 1, more than 50 SFO hypotheses at reader side are necessary for device with up to 10% SFO and 6 SFO hypotheses are sufficient at reader side for device with up to 1% SFO. 
For Option 3, 10 SFO hypotheses are sufficient for device with up to 10% SFO, but reader has to store the received samples and wait for the postamble that is after the end of PDRCH for any of SFO/channel/interference estimation, demodulation, decoding, etc.

For coherent detection of PDRCH with a payload of 400bits with 16-bit CRC, using 1/2 Manchester coding and 1/3 or 1/2 convolutional code, 
For option 1 of D2R preamble only, 
Sources [3, Huawei], [5, CMCC], [6, ZTE], [8, Spreadtrum], [9, vivo], [14, xiaomi] observed that with up to 10% SFO, 10% BLER cannot be achieved. 
Source [20, OPPO] observed that with perfect SFO estimation, 1kbps data rate and OOK modulation, there is no noticeable performance gain from using midamble(s) and/or postamble for PDRCH transmission.
For other amble options, 
Source [3, Huawei] observed that
With accurate SFO estimation, Option 2 of D2R preamble + 4 midambles can achieve 10% BLER at SNR ~ 2.7dB but cannot achieve 1% BLER.
With up to 10% SFO, Option 3 cannot achieve 10% BLER.
With up to 10% SFO, Option 4 of D2R preamble+2 midambles+postamble achieves 10% BLER at SNR of ~0.25dB; But it cannot achieve 1% BLER. Option 4 of D2R preamble+3 or 4 midambles+postamble, achieves a 10% BLER at an SNR of around -0.2 dB, and achieves 1% BLER at SNR around 9dB or 8dB, respectively.
Source [5, CMCC] observed that with up to 10% SFO, Option 4 of D2R preamble combined with 1 to 4 midambles + postamble, achieves 10% BLER at SNR of 2.5 dB, 1 dB, 0.8 dB, or 0.5 dB, respectively, for a data rate of around 1 kbps.
Source [6, ZTE] observed that with up to 10% SFO, 
Option 3 can provide ~5.5 dB performance gain compared to option 2 of D2R preamble+1midamble for 10% BLER, with the same amble(s) overhead for ~1kbps data rate.
Option 2 of D2R preamble+1midamble cannot achieve 1% BLER for ~1kbps data rate.
Option 4 of the D2R preamble+1 or 2 midamble(s)+postamble, has similar performance, it can achieve a 10% BLER at SNR of -1dB and achieves a 1% BLER at SNR of 6dB and 5dB respectively for ~1kbps data rate.  
Source [8, Spreadtrum] observed that with up to 10% SFO, 
Option 3 of D2R preamble+ postamble cannot achieve 10% BLER for ~7kpbs.  
Option 4 of D2R preamble + 1 midamble + postamble can achieve 10% BLER and 1% BLER at SNR around -6dB and 0 dB, respectively for ~7kpbs data rate.
Source [9, vivo] observed that 
With up to 10% SFO, maintaining the same amble overhead, both Option 2 of D2R preamble+5 midambles and Option 4 of D2R preamble+4 midambles+1 postamble demonstrate similar performance, achieving 10% BLER at SNR ~0.2dB. While for 1% BLER, the SNR for Option 2 and Option 4 is ~ 9.2dB and 12.8dB, respectively for ~5.5kpbs data rate.   
With up to 1% SFO, maintaining the same amble overhead, both Option 2 of D2R preamble+5 midambles and Option 4 of D2R preamble+4 midambles+1postamble demonstrate similar performance, achieving 10% BLER at SNR around -1.2dB. While for 1% BLER, the SNR for Option 2 and Option 4 is ~7.8dB and 9.1dB, respectively for ~5.5kpbs data rate.   
Source [14, xiaomi] observed that with up to 10% SFO, Option 2 of D2R preamble+3 midambles and Option 4 of D2R preamble+3 midambles+postamble can achieve 10% BLER when the SNR is within the range of [15, 25] dB for 1.25 kbps data rate. 
Source [30, Qualcomm] observed that the required SFO estimation accuracy to achieve 1% and 10% BLER is much smaller than 10^3 ppm. To achieve the required accuracy,
For Option 2 of D2R preamble+X midamble(s) where midamble inserted per every certain number of PDRCH bits (e.g., 192 bits),
For SFO estimation using each amble for the subsequent PDRCH bits (e.g., 192 bits), with up to 10% SFO, more than 50 SFO hypotheses are necessary at the reader side and with up to 1% SFO, 6 SFO hypotheses are sufficient at the reader side.
For SFO estimation based on the time gap between preamble and midamble, with up to 10% SFO, 10 SFO hypotheses are used, but reader has to store the received samples and wait for the midamble to start SFO/channel/interference estimation, demodulation, decoding, etc.
For Option 3 of D2R preamble+postamble, SFO estimation is based on the time gap between preamble and postamble, with up to 10% device SFO, 10 SFO hypotheses are used for reader, but reader has to store the received samples and wait for the postamble that is after the end of PDRCH for any of SFO/channel/interference estimation, demodulation, decoding, etc
Source [7, Samsung] observes that the additional ambles i.e., midamble(s) and/or postamble introduces additional overhead and postamble may prevent pipelined processing of the reception.

For the synchronization and timing tracking of D2R transmission, 
Source [5, CMCC] report that with up to 10% SFO, option 1 is not sufficient for D2R reception since the residual SFO at reader side is larger than 1%. While with option 3, the reader can precisely search and detect the SFO with a residual SFO of 0.03% at -3dB SNR TDL-A channel.
Source [14, xiaomi] report that 
For packet size of 96bits, when the SNR is increased from -4dB to 20dB, the ratio of device residual SFO over 100ppm decreases to 6% for Option 2, 3 and 4, but remains at 95% for Option 1.
For packet size of 400bits, when the SNR is increased from -4dB to 20dB, the ratio of device residual SFO larger than 10ppm decreases to 5% for Option 2, 3, and 4, but is still 99.6% for Option 1.
Sources [9, vivo], [15, CATT] report that SFO estimation based on D2R preamble can achieve accurate estimation without additional ambles (midamble or postamble). 
Source [9, vivo][7 Samsung] observed that for non-coherent detection of PDRCH, the number of SFO hypotheses and the SNR needed for 10% and 1% BLER cannot significantly be reduced for option 2, 3 and 4 compared to the option 1. Moreover, the additional ambles i.e., midamble(s) and/or postamble introduces additional overhead and postamble may prevents pipelined processing of the reception.     
Source [15, CATT] observed that 
The coarse estimation of SFO based on the D2R preamble indicates that the SFO estimation error is less than 1% with a probability of 99.3%, and less than 0.1% with a probability of 49.9%.
The fine estimation of SFO based on the D2R preamble shows that the SFO estimation error is less than 1% with a probability of 99.5%, and less than 0.1% with a probability of 90.8%.
Reader/gNB can achieve a probability of not less than 99.5% for SFO estimation error below 1%, and 90.8% for SFO estimation error below 0.1% by receiving D2R preamble signals.
Source [30, Qualcomm] report that for D2R with coherent demodulation at reader, the reader needs to estimate the device clock frequency with the accuracy of 0.5% (5 * 10^3 ppm) or lower for a short message (e.g., 72 bits after CRC/coding) and 0.1% (10^3 ppm) or lower for a long message (e.g., 224 bits after CRC/coding). The source further reports that design of D2R amble(s) (e.g., overhead) and the correspondingly required number of SFO hypothesis for the estimation depend on the sampling clock accuracy that the device uses for D2R. 
Source [37, MediaTek] reports that transmitting 96-bit packet size with 16-bit CRC requires residue SFO after reader compensation to be 1000 ppm, and transmitting 1000-bit packet size with 16-bit CRC requires residue SFO after reader compensation to be 100 ppm.
Note: in the observations above where coherent detection is used, sources that evaluated option 3 and option 4 assumed that the postamble is used at least for time/frequency tracking and for channel estimation.

Agreement
Following observations on R2D clock-acquisition part are captured in TR 38.769:
On impact/restriction of M values for the clock-acquisition part
9 sources [TCL, Nokia, Huawei, CMCC, ZTE, Apple, CATT, Mediatek, Qualcomm] provided observations on the impact/restriction of M values for the clock-acquisition part design requirements:
1 source [Nokia] observed that increasing value of M, while retaining the same transmission duration, improves the auto-/cross- correlation properties of the sequence due to increase in sequence length and use of   provides better timing estimation accuracy even in the presence of SFO as the sequence length spans only over a shorter duration. 
2 sources [TCL, Huawei] observed for option 1 of the clock-acquisition part design that no restriction is required to be placed on the M values. Furthermore, 1 source [Huawei] observed that the same 2 ON-OFF voltage (with the same duration) satisfies the FDR performance metric of less than 1% for different M values, e.g., M = 2, 6 and 24, where FDR is the False detection ratio (FDR), i.e. incorrectly calculating M, is the performance metric.
1 source [CMCC] observed that pattern of the clock-acquisition part is related to M chips per OFDM symbol and when M is small, the clock-acquisition part may cross multiple OFDM symbols, and the CP insertion may degrade the timing acquisition performance.
1 source [ZTE] observed that with option 2, the duration of the clock-acquisition part remains consistent across all M values, at least three OFDM symbols maybe required for clock-acquisition part and it maybe not as efficient as option1
1 source [Apple] observed that among the two options studied for the clock-acquisition part, option 2 provides increased robustness, especially in case of large value of M, when compared to option 1 and potentially increase the detection performance of the clock-acquisition part. 
1 source [CATT] observed that if the chip duration is variable based on the M value used for OOK-4 waveform, the detection performance would be limited by the received SINR of the CAP with clear transition of the rising and falling edges.
1 source [Qualcomm] observed that the option 1 with M>1 has shorter duration of clock acquisition part than M=1 and worse timing acquisition accuracy. At least part of PRDCH following the clock acquisition part may need to be used to improve the timing acquisition. Furthermore, the larger M (e.g., M>4) with small chip duration is more sensitive to the SFO accuracy and the restriction of M for the clock acquisition part may be needed.
1 source [Mediatek] further observed that different M values may impact the chip accuracy obtained by the clock acquisition part.

On impact of CP insertion/handling on the clock-acquisition part
10 sources [TCL, CMCC, ZTE, Samsung, Vivo, CATT, NTT Docomo, Qualcomm, Mediatek, Spreadtrum] observed that the CP insertion/handling may impact the design requirements of the clock-acquisition part:
1 source [CMCC] further observed that when the clock-acquisition part occupies more than one OFDM symbol, ON-OFF state transition around CP can avoid the error rising or falling edges due to the CP insertion.
1 source [ZTE] further observed that to mitigate the impact of the CP in the clock-acquisition part for large M values, it can reuse the CP handling method for PRDCH 
1 source [Samsung] further observed that CP insertion/handling on the clock-acquisition part can cause false rising/falling transition and, therefore, the clock acquisition part should be designed such that it does not incur a false rising or falling edges due to CP insertion when CP-OFDM is used for OOK signal generation.
1 source [vivo] further observed that CP insertion/handling on the clock acquisition part will impact the chip duration estimation accuracy. It is further observed that for CP handling, device may not be able to count the clock and estimate OFDM symbol duration accurately until the clock acquisition part if the start indicator only includes a single ON-OFF transmission. 
1 source [CATT] further observed that the SER will be degraded due to uneven chip interval when the CP is inserted within an OFDM symbol, where SER refers to the number of samples which is mismatched for comparing to the total number of samples in a chip.
1 source [NTT Docomo] further observed if CP insertion would cause false rising/falling edges, accuracy of timing acquisition may be impacted.   
1 source [Mediatek] further observed that the issues of chip extension, false raising/falling transition, and additional raising/falling transition caused by CP insertion/handling considering different M values will impact the chip accuracy obtained by the clock acquisition part.
1 source [Spreadtrum] further observed that the design of clock acquisition part should consider that CP insertion does not cause a false rising or falling edges and does not cause different length of multiple high / low voltages within the clock acquisition part when the clock acquisition spans multiple OFDM symbols.
1 source [Huawei] observed CP insertion/handling may not impact the design requirements of the clock-acquisition part

Agreement
For the D2R preamble design, following aspects have been studied and can be captured in the TR 38.769:
Autocorrelation Property
10 sources [Nokia, Huawei, CMCC, Xiaomi, CATT, Oppo, Ericsson, NTT Docomo, Qualcomm, ZTE] observed that the signal should have good autocorrelation properties for accurate peak detection based on the signal correlation at the reader 
Cross-correlation Property
7 sources [Nokia, CMCC, Oppo, Ericsson, Qualcomm, ZTE, CATT] observed that the signal should have good cross-correlation properties if multiple D2R preamble sequences are considered (e.g. for multiple access schemes (if supported) for D2R transmissions). 
Line coding
1 source [Nokia] observed that line coding may impact the autocorrelation property of the sequence. 
1 source [Huawei] observed that for D2R preamble, to apply backscattering, line coding can help improve the detection performance based on shifting the D2R signal’s frequency location away from the carrier wave
Sequence Types (not limited to below types only)
M-sequence
3 sources [Nokia, Vivo, Xiaomi] observed that m-sequence can be considered for D2R preamble mainly owing to good correlation properties.  
Golay sequence
4 sources [CMCC, Vivo, Xiaomi, Samsung] observed that Golay sequence can be considered for D2R preamble mainly owing to good correlation properties and availability of large number of distinct sequences and complementary pairs.  
Walsh sequence
1 source [Oppo] observed that Walsh sequence can be considered as a candidate for D2R preamble thanks to its good auto/cross-correlation property and flexible length
General Observations
1 source [Huawei] observed can achieve 0.97% residual SFO with 98% probability under -2.5dB SNR and 0.1% MDR with [-1/8, 1/8] chip timing error with 99.05% probability under -2.5dB SNR with D2R preamble including 2-parts with clock-like sampling frequency signal and timing-acquisition signal, having 32-length ‘1’ sequence (encoded to 64-chip Manchester code) and 32-length sequence (encoded to 64-chip Manchester code), respectively.
4 sources [TCL, CMCC, ZTE, Vivo] observed that for D2R preamble with binary signal, the timing synchronization performance is highly related to the sequence length of the preamble. Furthermore, 1 source [CMCC] observed that to achieve a BLER performance at 10%, the timing synchronization error should be less than 10%. Furthermore, 1 source [ZTE] observed that the channel estimation performance is also highly related to sequence length. 1 source [ZTE] observed that using a 32 bits preamble provides ~8 dB, ~5 dB performance gain than using 8 bits, 16 bits preamble, respectively. And using a 64 bits preamble provides ~2.5dB performance gain than using a 32 bits preamble.
1 source [Ericsson] observed that for D2R preamble with binary signal, normalized SFO estimation error of less than 10% can be achieved with a training sequence length 64 or longer. The simulated D2R preamble consisting of a Golay complementary pair can tolerate SFO up to 1% (AWGN) with up to 1 dB loss in performance for a sufficiently long preamble sequence length (32 or greater).

Agreement
For determining the end of PRDCH at the device, following two options are studied and captured in the TR 38.769:
Option 1: TBS information (via implicit/explicit L1 R2D control information)
Option 2: Postamble (at the end of PRDCH) 
14 sources [Nokia, Huawei, ZTE, CMCC, Samsung, Ericsson, Oppo, LGE, Qualcomm, Spreadtrum, Mediatek, Cewit, Ericsson, vivo] provided following observations on the above two options for determining the end of PRDCH:
3 sources [Nokia, Huawei, ZTE] observed that option 2 provide two benefits, namely, the variable payload length and to provide timing acquisition before the subsequent transmission of either PDRCH or PRDCH, thus improving the detectability at both reader and the device, respectively. Furthermore, 1 source [Huawei] observed that R2D postamble indicates the TBS with high efficiency for small packets by avoiding a large padding overhead, unlike option 1, which may require devices to perform blind detection of different PRDCH formats (if supported) and the overhead caused by the inclusion of a R2D postamble does not exceed 20% for even the smallest of message sizes and may be less than the signaling overhead caused by using a dedicated TBS indicator
1 source [CMCC] observed for option 2, that for small payload size with only a few bits, the presence of long postamble generates large resource overhead, while for large payload size with more bits, the resource overhead of postamble is smaller.
1 source [vivo] observed for option 2, that for small payload size with only a few bits, the presence of long postamble generates large resource overhead.
1 source [Samsung] observed option 2 is not strictly required, however, given the possible clock drift at a device, it may be still beneficial to also attach postamble at least for the determination of the end of PRDCH at a device. 
3 sources [Oppo, Spreadtrum, CEWiT] observed that with option 2, the false detection may be higher for shorter postamble. Source [OPPO[ observed that in contrast to option 2, it is more reliable and efficient to indicate TBS with control information in option 1
2 sources [LGE, vivo] observed that if a message type or a command ID is included in L1 control information and implicitly indicates a known size of a fixed TB, then there is no need for either option 1 or option 2
2 sources [Qualcomm, vivo] observed that option 1 has the advantages of avoiding blind detection of postamble and providing the power saving for non-target devices to skip the R2D detection.
1 source [MediaTek] observed that option 1 is feasible for the device to avoid the unnecessary reception of a TB with a specific size and thus enable power saving, e.g., when the TB has a size exceeding the allowance of the device remaining power.
1 source [Ericsson] observed option 2 is not strictly required if the end of PRDCH can be explicitly indicated by R2D control information, and it is subject to the miss-detection rate. It may be beneficial if a PRDCH postamble can serve as an additional timing acquisition signal prior to a PDRCH transmission.

Agreement
For D2R scheduling, midamble (if supported) related information can be explicitly/implicitly indicated via corresponding PRDCH.

Agreement
Following observations on R2D clock-acquisition part are additionally captured in TR 38.769:
On purpose of SFO estimation/correction based on the clock-acquisition part
3 sources [Nokia, CATT, Qualcomm] provided observations on the applicability of clock-acquisition part for frequency synchronization:
1 source [Nokia] observed that the length of preamble sequence may need to consider also the robustness against SFO
1 source [CATT] observed that device 2a/2b may require higher synchronization accuracy for signal transmission or backscattering and therefore, the design of CAP may be required to accommodate the requirement of additional frequency synchronization and clock calibration for Device 2a/2b. 
1 source [Qualcomm] further observed for Option 1, as the CAP duration with high M is decreased, only CAP may not be sufficient for SFO correction and for Option 2, as the CAP duration is fixed and independent from M, the CAP with long enough duration can support SFO correction.
On purpose of CFO estimation/correction based on the clock-acquisition part
2 sources [Ericsson, Qualcomm] provided observations on the applicability of clock-acquisition part for frequency synchronization:
1 source [Ericsson] observes that the clock-acquisition part can be utilized to solve the frequency synchronization problem without impacting the time-domain sequence, for example by transmitting in some frequency resources and it can be a harmonized solution for both chip duration indication and device frequency synchronization. However, it is further observed that if the time interval between an R2D transmission and the corresponding D2R transmission and if the device loses the timing obtained from the R2D timing acquisition signal due to timing drift at the time for the D2R transmission, then an additional synchronization signal is needed
1 source [Qualcomm] further observed for Option 1, as the CAP duration with high M is decreased, only CAP may not be sufficient for CFO correction and for Option 2, as the CAP duration is fixed and independent from M, the CAP with long enough duration can support CFO correction.

WI Phase: RAN1 Agreements (relevant for R2D/D2R signals including timing acquisition and synchronization
RAN1#120 (Athens, Greece, Feb 17th – 21st, 2025)
SIP related Agreements
Agreement
For the SIP of R-TAS, for providing the start of the R2D transmission, one single design based on Option 1 is supported and further down-selection to be done among Alt 1 and Alt 2 :
Option 1: ON-OFF transmission with following alternatives:
Alt 1: A single ON-OFF transmission with pre-defined duration for each of the ON-OFF, where ON and OFF may have same or different durations
Continue discussion to clarify the duration of each of the ON and OFF
Continue discussion to list the different candidate proposals under Alt1
Alt 2: A multi-ON-OFF transmission with pre-defined duration for each of the ON(s)-OFF(s), where different ON and different OFF may have same or different durations and different parts may have same or different duration
Continue discussion to clarify the duration of each of the ON and OFF
Continue discussion to list the different candidate proposals under Alt2
Only a single fixed value for entire duration of SIP of R-TAS is supported, which is independent of the value of “M” used in CAP and PRDCH
Note: Specific design and duration for SIP of R-TAS are further discussed, and companies are encouraged to evaluate the designs in terms of target MDR of [10%] for a FAR up to [1%] and at least following assumptions are used:
MDR refers to the probability that SIP is not detected when it was actually transmitted
FAR probability that the receiver incorrectly detects SIP when SIP was not transmitted
Energy/edge detection-based method is the baseline assumption for evaluation purpose
Continue discussion on necessary details for simulation assumptions

Agreement
For the SIP of R-TAS, down-select among the following candidates:
Alt 1 (Single ON-OFF transmission)
Alt 1-1: ON followed by OFF with same duration for both
Alt 1-2: ON followed by OFF with a duration ratio of 1:[2,3]
Alt 1-3: ON followed by OFF with a duration ratio of [2,3]:1
Alt 2 (Multi-ON-OFF transmission)
Alt 2-1: A number of repetition instances of Alt 1-1 or Alt 1-2 or Alt 1-3
Alt 2-2: ON-OFF-ON (duration of ON and OFF can be different)
Alt 2-3: OFF-ON-OFF (duration of ON and OFF can be different)
Alt 2-4: Combination of single instance of Alt 1-1 and single instance of Alt 1-2
For the evaluation purpose, for both options, candidate values related to duration are considered:
Entire duration of SIP: 1/2 OFDM symbol duration or 1 OFDM symbol duration (including clarifying whether OFDM symbol duration includes CP); additional durations can be considered and reported by companies with justification 
Companies to report the exact duration(s) for ON or OFF
Companies are encouraged to report at least the following details for the evaluations:
Baseline assumption is that RF transmission is not present; companies can report other consideration
For FAR calculation, whether noise and/or PRDCH transmission is considered
Details on threshold detection method including whether/how threshold detection training is used based on the proposed design alternative or not
BW assumptions for RF-ED and BB-LPF
Target MDR of up to 1% for FAR of up to [1%, 10%]

CAP related Agreements
Agreement
For the CAP of R-TAS, the starting chip has a different voltage level compared to the end of the SIP of R-TAS.

Agreement
For the design of the CAP of R-TAS, at least 2 transition edges in same direction are included, i.e. at least two transitions from “OFF” chip to “ON” chip or two transitions from “ON” chip to “OFF” chip.

Agreement
For the CAP of R-TAS:
Candidate values for maximum duration of CAP to be further down-selected to one value from : 1.5 OFDM symbol duration, 2 OFDM symbol duration, 3 OFDM symbol duration
For option 1 for CAP of R-TAS from TR 38.769, maximum duration is applicable to minimum value of M to be supported, and the CAP duration becomes shorter with increasing value of M
FFS: whether the number of ON/OFF transmissions in the CAP is fixed or not fixed
For option 2 for CAP of R-TAS from TR 38.769, maximum duration is the only (constant) duration that is applicable for all the M values to be supported
Down-selection between option 1 and option 2 for CAP of R-TAS from TR 38.769 by RAN1#120-bis
FFS: Values of M to be supported


R2D Midamble related Agreement
Agreement
R2D transmission does not include a midamble.

D2R X-amables related Agreement
Agreement
For D2R preamble design, the functionalities of timing acquisition, SFO estimation/time tracking and channel estimation should be supported
For D2R midamble design, the functionalities of SFO estimation/time tracking and channel estimation should be supported
D2R midamble can be transmitted at the end of the PDRCH transmission. If it is at the end, it is not designed for being used for indicating the end of PDRCH transmission
FFS: condition(s) and/or indication where the D2R midamble is present or not

Agreement
For D2R x-ambles:
Following is considered as the types for base sequence and to be further down-selected:
Option 1: M-sequence 
Option 2: Golay sequence
Note: Above doesn’t preclude an additional part for preamble, e.g. with ON and/or OFF transmission, if needed/supported
FFS: Whether/what multiple sequences (using same base sequence type) are supported
Note: This in no way implies that there is going to be CDMA between D2R x-ambles
For evaluation purpose, companies are encouraged to consider following:
Performance at least in terms of autocorrelation/cross-correlation property, SFO estimation/Timing accuracy, SNR for target PDRCH BLER of [1%, 10%]
Report presence and time-domain resource(s) x-ambles
Report sequence type(s) and length(s) for x-ambles
Following format can be considered for reporting the evaluation results

RAN1#120bis (Wuhan, China, April 7th – 11th, 2025)
R-TAS related Agreements (including SIP and CAP)

Agreement
For the pattern of SIP of R-TAS, only the following 2 alternatives are considered for further down-selection:
Alt 1-2: ON-OFF with a ratio of 1:3 and with following total SIP duration to be further down-selected:
Option 1: 0.5 OFDM symbol duration
Option 2: 1 OFDM symbol duration
Alt 2-4: ON-OFF-ON-OFF with a ratio of 1:1:1:3 and with following total SIP duration to be further down-selected:
Option 1: 0.5 OFDM symbol duration
Option 2: 1 OFDM symbol duration

Agreement
For CAP of R-TAS, following is adopted:
Option 1 for CAP of R-TAS from TR 38.769 is adopted where the CAP duration becomes proportionally shorter with increasing value of M, i.e. if for , duration is  OFDM symbol long, then for , duration is  OFDM symbol long
Note: Duration without CP insertion is considered above, with CP insertion, the total duration may not be exactly proportional
Only following two alternatives for CAP pattern are considered for further down-selection to one alternative:
Alt 1: ON-OFF-ON-OFF
Alt 2: ON-OFF-ON


Agreement
For R-TAS, SIP duration of 1 OFDM symbol is adopted with CAP pattern ON-OFF-ON-OFF for all values of M corresponding to PRDCH 
Note: device cannot assume the presence/absence of RF transmission prior to the SIP.

D2R x-ambles related Agreements (including preamble and midamble)

Agreement
For D2R midamble, for determining the presence and location of midamble(s) at the device:
Reader explicitly indicates the same interval between consecutive midambles, and between the preamble and the first midamble, via R2D control information
FFS: details of signalling
FFS: whether the reader can explicitly indicate with one bit whether a midamble is additionally present at the end
Note: This does not preclude the support of having no midamble present in the D2R transmission

Agreement
For D2R preamble/midamble, base sequence is generated from m-sequence, where the length of the sequence is 
Value(s) of n
Long preamble/midamble is generated based on n = 5
Working assumption: Short preamble/midamble is generated based on n=3 
Only 1-part preamble/midamble are supported for D2R
Preamble immediately precedes the PDRCH without any gap
Both long and short preamble and midamble are supported based on the working assumption on n
when midamble is present at least the following cases are supported and reader explicitly indicates one of the following cases for PDRCH:
Short preamble and short midamble 
Long preamble and long midamble 
Note: the case of short preamble and long midamble will not be supported
When midamble is not present the reader explicitly indicates short or long preamble for PDRCH

Agreement
For indicating the interval between consecutive midambles, and between the preamble and the first midamble, via R2D control information, following is adopted:
Unit of interval is number of bits after FEC (if FEC is applied) and repetition (if repetition is applied)
FFS: the candidate values in terms of the unit of interval

TDoc file conclusion not found