R1-2601837
discussion
On transmission schemes for uplink channels in 6GR
From Nokia
Nokia's prior position on
10.5.2.3
at
RAN1#124
· AI-synthesized, paraphrased
verify sources →
Proposes simplifying 6GR UCI arbitration by limiting CSI reporting to PUSCH only and adopting slot-level one-shot multiplexing/prioritization to eliminate iterative symbol-level procedures of NR. Argues NR's five PUCCH formats are excessive and proposes a limited set of scalable PUCCH formats with high similarity to PUSCH design for larger UCI sizes. For PUSCH, identifies MIMO layer imbalance as a significant performance degradation factor and proposes studying multi-codeword operation and MCS table optimizations to decouple modulation order from code rate, along with new codeword-to-layer mapping to mitigate imbalance. Requires a single unified DMRS type with a scalable comb-type pattern supporting up to 48 orthogonal antenna ports, using Rel-18 e-type 1 as baseline, and proposes native multi-slot PUSCH resource allocations exceeding 14 symbols rather than NR's slot-bounded framework. For AI/ML receivers, presents simulation evidence of throughput gains with sparse DMRS and proposes studying three categories of sparse DMRS patterns (time-domain, frequency-domain, and combined) within the unified 6G DMRS framework while emphasizing realistic aligned evaluation assumptions for both AI and legacy receivers.
Summary
This 3GPP Tdoc from Nokia for the FS_6G_Radio work item provides a comprehensive analysis and set of 21 observations and 40+ proposals on uplink physical channel design for 6G Radio. It covers PUCCH simplification, PUSCH transmission schemes (codebook/non-codebook, full-power, multi-TRP), codeword-to-layer mapping for layer imbalance, DMRS design, resource allocation, power control, and AI/ML-based DMRS overhead reduction.
Position
Nokia proposes a streamlined 6G uplink control channel design by limiting CSI reporting to PUSCH and using PUCCH slot-level UCI multiplexing to avoid NR's complex iterative UCI arbitration. For the data channel, Nokia proposes SRS-based codebook-based PUSCH as the baseline, with studies on a unified precoding framework supporting per-port amplitude scaling and dynamic coherency conditions, while deprioritizing non-transparent open-loop and multi-TRP PUSCH schemes due to limited 5G adoption and implementation complexity. Nokia provides evidence that multi-codeword transmission from rank 2 with SIC receivers can achieve up to ~30% throughput gain over single-codeword mapping under layer imbalance. The company requires a single DMRS type (comb-based pattern) supporting up to 48 orthogonal ports, proposes multi-slot PUSCH TDRA exceeding 14 symbols, and advocates for contiguous-only FDRA Type 1, while emphasizing the need for fairness in AI/ML receiver evaluations using sparse DMRS.
Key proposals
- Proposal 1 (Sec 2.1): Study limiting CSI reporting to PUSCH for 6GR.
- Proposal 2 (Sec 2.2): In the design of 6GR PUCCH with a limited set of PUCCH formats, consider PUCCH formats that are scalable in terms of supported UCI sizes and PUCCH durations.
- Proposal 1 (Sec 3.1.1): The supported number of UE TX antenna ports for 6GR are {1, 2, 3, 4, 8}.
- Proposal 7 (Sec 3.1.2): Study a unified UL precoding framework for all UE device types and antenna layouts focusing on: Uplink CSI feedback design from aNB to UE to enable the use of amplitude and phase values. Scalable phase and amplitude granularity to control trade-off between performance and overhead.
- Proposal 1 (Sec 3.2): In 6G, codeword‑to‑layer mapping schemes enabling a second codeword from rank 2 (up to 4 CWs) should be studied for uplink considering rank‑adaptive operation, different traffic models, and receiver capabilities (e.g., MMSE‑IRC and SIC).
- Proposal 1 (Sec 3.3.1): 6GR should strive for single DM-RS type and single scalable pattern considering channel estimation performance and complexity.
- Proposal 2 (Sec 3.3.1): Consider Rel-18 e-type 1 DMRS (UL) as a baseline for possible 6GR DMRS candidates.
- Proposal 3 (Sec 3.3.1): Support only up to 48 orthogonal DMRS antenna ports.
- Proposal 1 (Sec 3.4.1): Study PUSCH resource allocations of more than 14 OFDM symbols in addition to the NR (slot-based) PUSCH (repetitions).
- Proposal 1 (Sec 3.4.2): 6GR to support only contiguous PUSCH frequency domain resource allocation based on FDRA Type 1.
- Proposal 2 (Sec 3.5.1): 6G specifications should be simplified by adopting the (5G) Rel-17+ unified TCI principles as baseline for UL PC, i.e., the support for the Rel-15 spatial relation framework should be dropped in 6G.
- Proposal 1 (Sec 4.2): RAN1 defines the evaluation assumptions for AI-receivers (e.g. benchmark definition, DMRS configuration, and use of freed-up REs) to compare DMRS patterns (including sparse DMRS) for legacy receivers and AI.
- Proposal 1 (Sec 4.3): For AI receivers, RAN1 to study sparse DMRS configurations for 6G that reduce DMRS overhead while maintaining compatibility with the unified 6G DMRS framework.