R1-2508687
discussion
Discussion on modulation for 6GR air interface
From Xiaomi
Summary
This document contains 11 proposals and 9 observations from Xiaomi on 6G Radio (6GR) modulation. It addresses reusing NR uniform QAM schemes as baseline, studying higher-order 4096QAM/1024QAM with specific evaluation criteria, evaluating new constellation shaping schemes like NUC and PS-QAM, reusing NR's BICM framework, and studying a more flexible MCS table mechanism. The proposals span across NR baseline modulation, maximum modulation order, potential new modulation schemes, joint coding and modulation, and interim milestones.
Position
Xiaomi supports NR uniform QPSK, 16QAM, 64QAM, 256QAM, and 1024QAM as the basis for 6GR data channels, arguing there is no reason not to support NR scheme in the SI phase. For higher modulation orders, Xiaomi proposes studying uniform 4096QAM for DL and 1024QAM for UL, requiring Net Gain evaluation criteria (SNR degradation minus PAPR degradation) with a realistic PA model and RAN4 early involvement. Xiaomi presents a technical case against GS schemes like 2D-NUC, providing LLS results showing that ATSC 3.0 2D-NUC 256QAM provides only marginal gain (0.1~0.4 dB) in TDL-A channels with 48 PRBs and practical channel estimation, and argues the motivation for GS is not well justified given additional complexity and PAPR increases. On joint coding and modulation, Xiaomi proposes reusing the NR BICM framework and studying a more flexible MCS table mechanism to address the increasingly complex MCS table selection in NR.
Key proposals
- Proposal 1 (Sec 2.1): Supports 5G NR uniform QPSK, 16QAM, 64QAM, 256QAM, and 1024QAM as basis for 6GR DL data channel, and corresponding UL modulations for CP-OFDM and DFT-s-OFDM.
- Proposal 4 (Sec 2.2): For studying uniform 4096QAM for DL and 1024QAM for UL, requires using Net Gain [dB] = SNR degradation/gain relative to reference @10% BLER - PAPR degradation as the evaluation criterion, with a realistic PA model.
- Proposal 5 (Sec 2.2): For studying uniform 4096QAM for DL and 1024QAM for UL, requires studying associated restrictions including applicable frequency ranges, device type(s), number of MIMO layers, scaling factor to maximum data rate.
- Proposal 7 (Sec 3.1): For evaluation purposes, proposes using the ATSC 3.0 scheme as the candidate GS (Geometric Shaping) scheme.
- Proposal 8 (Sec 3.3): Requires RAN1 to first establish a clearer and aligned understanding of the proposed constellation shaping schemes, including design principles, advantages, limitations, and potential specification impacts, before collecting simulation results.
- Proposal 9 (Sec 4.1): Proposes reusing the 5G NR BICM framework in 6GR for coding-modulation concatenation.
- Proposal 10 (Sec 4.2): Proposes studying a more flexible mechanism for defining and selecting MCS tables for 6GR.
- Proposal 11 (Sec 5): For the interim milestone, proposes aiming for an earlier decision before June 2026, deciding whether to support enhancements beyond NR baseline, and clearly defining their scope to avoid hardware impacts.
- Proposal 2 (Sec 2.1): For channels other than data channel, requires RAN1 to clarify under which agenda item the discussion should proceed.
- Proposal 3 (Sec 2.1): For enhanced modulation schemes for PAPR reduction based on uniform modulation, requires RAN1 to clarify under which agenda item the discussion should proceed.
- Proposal 6 (Sec 2.2): Requires early involvement of RAN4 for the study of uniform 4096QAM for DL and uniform 1024QAM for UL.
- Observation 5 (Sec 3.3): For 256QAM in AWGN channel, the 2D-NUC scheme in ATSC 3.0 can provide about 0.42~0.77 dB gain relative to NR baseline at BLER 10%.
- Observation 6 (Sec 3.3): For 256QAM in TDL-A channel, the 2D-NUC scheme in ATSC 3.0 can provide about 0.49~0.8 dB gain relative to NR baseline at BLER 10% with 6 PRBs.
- Observation 7 (Sec 3.3): For 256QAM in TDL-A channel, the 2D-NUC scheme in ATSC 3.0 provides very marginal gain (almost no gain) relative to NR baseline at BLER 10% with 48 PRBs.