R1-2409668
discussion
Specification support for beam management
From vivo
Summary
This document from vivo addresses specification support for AI/ML-based beam management in NR, focusing on consistency issues, performance monitoring, and reporting overhead reduction. It contains 53 proposals and 8 observations covering UE-side and NW-side models, associated ID interactions, Set A/B configurations, and beam failure detection enhancements.
Position
Vivo argues that ensuring consistency of Set A and Set B properties (beam width, pointing angle, indexing) via an associated ID is critical to prevent severe performance degradation, prioritizing local associated IDs to protect NW proprietary information. They oppose Option 1 for associated ID interaction due to signaling waste and specification impact, preferring Options 2 and 3 where associated IDs are interacted before inference configuration. For performance monitoring, Vivo supports L1 reporting and UE-assisted monitoring to reduce UCI overhead, specifically recommending dedicated CSI-ReportConfigs for monitoring rather than embedding them in inference configs. Regarding resource configuration, they propose allowing UEs to recommend Set B patterns and using quasi-best Rx beams derived from limited P3 measurements to balance performance and overhead. For reporting, Vivo supports Time Resource Indicators (TRI) and differential quantization with a single reference value to compress BM-Case 2 reports, and suggests reusing existing L1-RSRP reporting structures for NW-side data collection while enhancing quantization steps.
Key proposals
- Proposal 1 (Sec 2.1): Support providing an associated ID representing NW-side additional conditions to the UE to ensure consistency between training and inference.
- Proposal 6 (Sec 2.4): Prioritize Option 2 and Option 3 in the applicable functionality determination procedure for interacting associated IDs before inference configuration.
- Proposal 11 (Sec 3): Support L1 beam reporting for performance monitoring for both UE-side and NW-side models.
- Proposal 19 (Sec 4.1): Support UE recommending preferred Set B patterns trained during the UE-side model training phase.
- Proposal 25 (Sec 4.5): Support using a quasi-best Rx beam for Set A and/or Set B measurement, derived from P3 measurement on a small number of Tx beams.
- Proposal 29 (Sec 6.1): Support reporting predicted L1-RSRPs and corresponding beam information of up to M beams within X dB gap to the largest predicted value.
- Proposal 36 (Sec 6.4): Support reporting TRI (Time Resource Indicator) instead of direct predicted beam resource indication to reduce overhead for BM-Case 2.
- Proposal 42 (Sec 7.1): Support reusing current specification report content (L1-RSRPs and beam information) for NW-side model data collection.
- Proposal 49 (Sec 7.4): Support enhancements on quantization range and step to reduce overhead for NW-side model measurement reports.
- Proposal 52 (Sec 9): Support initiating the data collection procedure for UE-side models via either gNB configuration or UE request signaling.
- Proposal 53 (Sec 10): Support using AI beam prediction for beam failure detection/report enhancement.
- Proposal 15 (Sec 3.3): Support dedicated resource sets and report configurations for monitoring in a dedicated CSI report configuration (Option 2) for Type 1 Option 2.
- Proposal 22 (Sec 4.2): Indicate Set A or resources in Set A as virtual sets/resources for inference, which do not require measurement.
- Proposal 47 (Sec 7.2): Introduce a beam index type indicator to specify if reported beams are highest or lowest quality for NW-side inference.