R1-2601812 discussion

Discussion on uplink transmission scheme(s) for uplink channels for 6GR

From Spreadtrum
Status: not treated
WI: FS_6G_Radio
Agenda: 10.5.2.3
Release: Rel-20
Source: 3gpp.org ↗
Spreadtrum's prior position on 10.5.2.3 at RAN1#124 · AI-synthesized, paraphrased
verify sources →
Proposes using NR closed-loop transmission (codebook-based and non-codebook-based) as the baseline for 6G PUSCH, but requires limiting 8-port transmission to FFS status for CPE/FWA only due to lack of commercial 8-port devices despite NR spec support. Proposes studying a simplified unified uplink codebook with antenna port group configuration to cover full-coherent/partial-coherent/non-coherent types, full power mode transmission, and asymmetric UL panels simultaneously. Requires defining a single DMRS type for both DL and UL to avoid implementation complexity from multi-type support, and supports DMRS overhead reduction sub-case A (sparse orthogonal DMRS in frequency and/or time domain) as a 6GR Day 1 AI/ML use case backed by simulation results showing AI channel estimation with 2-symbol DMRS achieving similar BLER to non-AI with 3-symbol DMRS. Opposes prioritizing superimposed pilot (SIP) as a use case pending further multi-user interference evaluation and standardization impact analysis. For PUCCH, proposes radical simplification from five NR formats to two formats (≤2 bits and >2 bits), eliminating sub-slot based PUCCH, two PUCCH configurations, semi-persistent PUCCH, and PUCCH SCell, while retaining PUCCH group concept and supporting proactive conflict avoidance for PUCCH resource determination.

Summary

Spreadtrum's contribution to 3GPP RAN1#124b provides 23 proposals and 5 observations on 6G uplink transmission schemes, covering PUSCH antenna ports, DMRS design with AI/ML overhead reduction, resource allocation, PUCCH formats, and new technologies like L1 UE aggregation and remote shared uplink panels.

Position

Spreadtrum proposes limiting 6G Day 1 PUSCH antenna ports to {1,2,4}, with 8-port studied only for CPE/FWA, arguing 8-port UEs have not been commercially used despite NR spec support. They propose a simplified unified uplink codebook supporting coherent type consumption, full power transmission, and symmetric/asymmetric UL panels, and study of frequency selective precoding for PUSCH. For multi-TRP, they require using NR schemes as baseline while avoiding multiple transmission schemes for similar scenarios, and propose studying L1 UE aggregation and remote shared uplink panels for performance enhancement. They propose a single DMRS type for UL supporting up to 32 ports with FDM/TDM/CDM multiplexing, and present simulation results showing AI channel estimation with 2-symbol DMRS achieves similar BLER to non-AI with 3-symbol DMRS, supporting DMRS overhead reduction sub-case A for Day 1 AI/ML. They take a cautious stance on superimposed pilot, requiring sufficient evaluation of multi-user interference and standardization impact before selection as priority use case.

Key proposals

Your notes

Private to your account