R1-2508730
discussion
Discussion on modulation, joint channel coding and modulation for 6GR
From OPPO
Summary
OPPO's contribution (R1-2508730) presents 3 formal proposals and 18 observations regarding 6GR modulation schemes, evaluating higher-order modulation, constellation shaping (geometric and probabilistic), and joint channel coding and modulation. The document argues against 4096QAM due to stringent EVM requirements and limited applicability, supports geometric shaping (GS) over probabilistic shaping (PS) due to PS's performance losses in fading channels and high complexity, and suggests further study on enhanced AMC and interleaver design.
Position
OPPO proposes capping DL modulation at 1024QAM and UL at 256QAM, presenting a technical case against 4096QAM based on a -38 dB EVM requirement causing deployment barriers and marginal system throughput gains (top 1% UEs in UMa/UMi at 7 GHz). They support geometrical shaping (GS) for 6GR, specifically recommending further study of 1D-NUC and 2D-NUC constellations, while presenting extensive technical evidence against probabilistic shaping (PS) due to performance loss in i.i.d. Rayleigh fading (0.4-0.7 dB loss), small resource allocations (up to 3.1 dB loss with 50 REs), and multi-layer MIMO transmissions (up to 1.0 dB loss for 2-layer). OPPO requires further study of joint channel coding and modulation enhancements (enhanced AMC with overlapped MCS) contingent on integration with UE CQI feedback procedure, and argues that customized interleavers for unbalanced bit-level reliability introduce additional implementation difficulties.
Key proposals
- Proposal 1 (Sec 2.1): Up to 1024-ary constellation for DL and up to 256-ary constellation for UL are supported for 6GR modulation scheme.
- Proposal 2 (Sec 2.2.3): Support geometrical shaping for 6GR modulation scheme.
- Proposal 3 (Sec 2.2.3): For geometrical shaping, further study 1D-NUC and/or 2D-NUC for 6GR modulation scheme.