R1-2601877
discussion
Considerations for the downlink control channel, scheduling for downlink and uplink transmissions
From Kyocera Corporation
Summary
Kyocera presents 5 Observations and 4 Proposals arguing for a single-stage DCI design with a fixed-size unified format for 6G, moving feature-dependent fields from PDCCH to PDSCH or RRC to limit blind decoding hypotheses and candidates, thereby reducing UE energy consumption without a two-stage DCI architecture.
Position
Kyocera proposes a single-stage DCI baseline for 6G, explicitly arguing that two-stage DCI introduces multi-step dependencies, timing constraints, and additional failure modes which increase latency and implementation complexity. They propose a fixed-size unified DCI structure (40-bit example) for both DL and UL scheduling that separates baseline DCI fields on PDCCH from feature-dependent information, moving the latter to PDSCH-embedded signaling for DL and RRC-configured profile-based completion for UL. They propose hash-conditioned early termination using a Target Identification Field (TIF) mapped to early polar bit positions, allowing non-target UEs to abort decoding after partial polar processing. They propose restricting SearchSpaceSets, aggregation levels, and candidate counts as intentional configuration levers to bound blind decoding complexity, and they propose handling cross-carrier scheduling either via SearchSpace-bound carrier context or a compact Carrier-Group ID (1–2 bits) without increasing DCI size hypotheses.
Key proposals
- Proposal 1 (Introduction): RAN1 should consider adopting single-stage DCI as the baseline design for the 6GR study.
- Proposal 2 (Fixed-size unified DCI structure): RAN1 should review all DCI fields for assignment to either baseline or feature-dependent fields.
- Proposal 3 (Fixed-size unified DCI structure): RAN1 should consider various enhancements, such as implementing a fixed-size baseline DCI, when studying single-stage vs. two-stage DCI.
- Proposal 4 (UL scheduling): RAN1 should consider various design options to enhance the single-stage DCI design, such as piggybacking PDSCH and/or higher-layer feature-dependent DCI fields.