R1-2508738 discussion

Modulation for 6GR air interface

From Huawei
Status: noted
WI: FS_6G_Radio
Agenda: 11.4.2
Release: Rel-20
Source: 3gpp.org ↗

Summary

Huawei presents a comprehensive technical case against adopting constellation shaping (both geometric and probabilistic) for 6G, documenting 73 observations and 13 proposals. The document argues that shaping gains observed in AWGN channels diminish or reverse in fading channels, while complexity (up to 100x LDPC decoding), throughput bottlenecks (~10x processing delay), and PAPR increases make shaping impractical; instead, Huawei proposes enhanced adaptive modulation and coding (E-AMC) which achieves up to 1.5dB gain with negligible additional complexity.

Position

Huawei presents a detailed technical case against introducing constellation shaping (geometric and probabilistic) in 6G, arguing that AWGN-optimized shaping parameters suffer substantial gain reduction or become losses in fading channels (e.g., CCDM loses 0.5dB in SISO and 0.8dB in rank-4 MIMO). They propose studying enhanced adaptive modulation and coding (E-AMC) as a superior alternative, demonstrating up to 1.5dB gain in fading conditions by jointly optimizing modulation order and code rate based on channel characteristics with almost no additional complexity. Huawei documents prohibitive complexity overheads: 2D-NUC demodulation is 14.4x LDPC decoding for 1024QAM, CCDM incurs ~10x processing delay over LDPC decoding, and reduced ML detection with 2D-NUC is 10-100x LDPC decoding complexity. They require that performance results in future 6G modulation studies must be reported together with transmitter/receiver complexity, latency impact, parallelism implementation, and storage requirements, and stress that RAN4 confirmation is needed for phase noise, EVM requirement, MPR/A-MPR under realistic PA models. Huawei also observes that both geometric and probabilistic shaping increase PAPR for DFT-s-OFDM, counteracting its power efficiency benefit.

Key proposals

Your notes

Private to your account