R1-2508981
discussion
FL summary#1 for 6G channel coding
From ZTE
Summary
This document is the Feature Lead (FL) summary from ZTE and Apple for 3GPP RAN1#123 on 6G channel coding. It compiles 22 distinct proposals and observations from 30+ contributing sources, covering data channel LDPC enhancements for higher throughput, control channel Polar code extensions for larger payloads, and evaluation methodologies for both.
Position
ZTE serves as moderator alongside Apple, synthesizing 30+ company contributions without taking explicit positions on most technical options. In the data channel section, ZTE's own contributions project target peak data rates up to 8x NR and observe that for fast-convergence LDPC with 2 iterations, new BG design achieves over 1 dB gain versus NR BG1 at code rates above 2/3. ZTE provides specific computational complexity formulas (C_total = I*M(2dc+2) additions + I*M(2dc-3) comparisons for LMS decoding) and observes that hardware complexity of increasing lifting size is higher than using multiple decoders, degrading decoding efficiency. For control channel, ZTE observes performance degradation even within NR UCI range due to excessive repetition and reports 0.35-2.07 dB BLER gains by increasing segments for 1024-1706 bit payloads at code rate 1/12. ZTE proposes Packet coding and intra-CBG interleaving for data channel chain enhancements, reporting 0.5 dB PER gain and 0.5-7 dB performance gains respectively. The document frames proposals as FL observations rather than company positions, with ZTE's technical input appearing in the detailed observation sections.
Key proposals
- Proposal 3.1.1-1-v1 (Sec 3.1.1): For the evaluation of 6G data channel coding scheme(s) for higher throughput with acceptable performance-complexity tradeoff, at least the following metrics should be reported: BLER performance, Complexity, Decoding throughput/latency. Area efficiency and Energy efficiency can be also reported.
- Proposal 3.1.2-1-v1 (Sec 3.1.2): For the study of 6G data channel coding scheme(s), specifies evaluation assumptions including AWGN channel, QPSK baseline modulation (16QAM/64QAM/256QAM also reportable), layered BP decoding with 2-20 iterations and reversed decoding order.
- Proposal 3.2-1-v1 (Sec 3.2): For LDPC code for higher throughput, six options are identified: (1) Reduce maximum number of iterations/fast convergence, (2) Increase maximum lifting size to [2 or 4]*384, (3) Optimize parallelism/orthogonality between rows, (4) Increase number of systematic columns, (5) Reduce number of edges in BG, (6) Implementation based solutions. QC-LDPC structure is confirmed, with FFS on using 5G BG(s) or defining new BG(s).
- Proposal 3.2-3-v1 (Sec 3.2): Study whether/how to optimize the error floor performance for LDPC code, targeting scenarios such as URLLC and NTN service.
- Proposal 3.2-4-v1 (Sec 3.2): Study whether/how to enhance LDPC for 6G low complexity/power scenarios, such as IoT/RedCap UEs, including compressing base graph or reducing maximum lifting size.
- Proposal 3.3-1-v1 (Sec 3.3): For 6G data channel coding chain enhancements, proponents should provide details on inter-code block coding and bit interleaver enhancement, evaluation metrics including BLER/throughput, and evaluation assumptions covering channel type, resource allocation, MCS, channel estimation, and Tx/Rx antenna configuration.
- Observation 4.1-1-v1 (Sec 4.1): [20 sources] discussed DCI channel coding within NR range (≤140 bits). [16 sources] suggested using NR Polar code. Additional discussions covered early termination (D-CRC, PAC code, new integrity check mechanism), RNTI FAR issues, higher modulation order (MLC framework showing 0.2-0.3 dB gain over BICM for 16QAM), and CRC overhead reduction.
- Observation 4.1-2-v1 (Sec 4.1): [12 sources] discussed DCI beyond NR range. Necessity debated (5 sources see need, 5 sources dependency, 2 unclear). Coding options include segmentation, D-CRC interleaver removal/redefinition, PAC codes (60-65% TSCCR vs <30% for NR D-CRC), 2-stage DCI, and 1024-length Polar sequence for DL (0.3-1.1 dB gain observed).
- Observation 4.2-1-v1 (Sec 4.2): [21 sources] discussed UCI within NR range (≤1706 bits). [16 sources] suggested NR Polar code, but [3 sources] observed BLER degradation with NR segmentation. Proposed solutions include new segmentation schemes (up to 2 dB gain), TPAC codes (up to 2.2 dB gain), and higher modulation order.
- Observation 4.2-2-v1 (Sec 4.2): [14 sources] discussed UCI beyond NR range (>1706 bits). [9 sources] suggested more than 2 segments (showing 2-5 dB gain for payloads 1706-2026 bits). Other options include TPAC codes, N/2 Polar code sequence (0.2 dB gain), and higher modulation order.
- Proposal 4.3-1-v1 (Sec 4.3): For 6G small UCI channel coding, study 5G RM code reuse, enhanced coding schemes (including rate matching), and use of sequences, with FFS on range (e.g., 3-11 bits). Enhanced schemes show 0.7-1.4 dB gain over RM code at higher code rates.
- Observation 4.4-1-v1 (Sec 4.4): [17 sources] discussed PBCH channel coding. [16 sources] suggested NR Polar code. One source observed that reducing CRC from 24 bits to 16 bits may provide 1 dB coverage improvement.
- Proposal 4.5-1-v1 (Sec 4.5): For 6G control channel coding chain enhancements, proponents should provide details and spec impact, evaluation metrics (BLER, throughput), and evaluation assumptions including channel type, resource allocation, MCS, channel estimation method, and Tx/Rx antenna configuration.
- Proposal 4.6.1-1-v1 (Sec 4.6.1): For 6G control channel coding evaluation, at least the following metrics are considered: BLER performance, FAR, Complexity, Early termination, and DCI RNTI false alarm.
- Proposal 4.6.2-1-v1 (Sec 4.6.2): Evaluation assumptions for 6G control channel coding include AWGN channel, QPSK modulation (other orders reportable), SCL decoding with list size 8, UL code rates from 1/12 to 5/6, and DL coded bit lengths of 108*[1,2,4,8,16].