R1-2410504
discussion
Specification support for beam management
From KDDI Corporation
Summary
KDDI Corporation proposes 7 specifications for AI/ML beam management in 5G NR, focusing on overhead reduction mechanisms for network-sided models and performance monitoring approaches for UE-sided models. The document addresses both temporal correlation exploitation in BM-Case2 and efficient monitoring alternatives for beam prediction accuracy.
Position
KDDI advocates FOR overhead reduction mechanisms that exploit temporal correlation in RSRP values and flexible threshold-based beam reporting, while pushing FOR efficient Alt 3 performance monitoring over more resource-intensive alternatives. They advocate AGAINST approaches requiring full Set A measurements and express skepticism about Alt 4 probability-based monitoring, questioning its practical value over simpler alternatives.
Key proposals
- Proposal 1 (Sec 1.1): For NW-sided model in BM-Case2, introduce overhead reduction mechanism by utilizing temporal correlation of RSRP values across multiple time instances
- Proposal 2 (Sec 1.2): For NW-sided model inference report, support reporting of L1-RSRPs and corresponding beam information of up to M beams within X dB gap to the largest measured value of L1-RSRP, where X and M are configured by gNB
- Proposal 3 (Sec 1.2): For NW-sided model, support mechanism for UE to report the number of beams it intends to report, either periodically or based on network-specified events, to enable adaptive beam reporting with optimal overhead-performance trade-off
- Proposal 4 (Sec 2.1): For UE-sided model performance monitoring under Alt 1 and Alt 2, Full measurement of Set A beams is required to obtain ground truth information about the correct Top 1/Top K beams for calculating beam prediction accuracy
- Proposal 5 (Sec 2.1): For UE-sided model performance monitoring, support Alt 3 as it enables efficient performance monitoring by measuring only the beams of the monitoring resources predicted by AI/ML based on Set B measurements
- Proposal 6 (Sec 2.1): When implementing Alt 3, if the monitoring resources are configured to match the Top 1/Top K beams, the system can achieve efficient monitoring by utilizing information obtained during the inference operation
- Proposal 7 (Sec 2.1): Further discussion is needed regarding the use of probability information in monitoring under Alt 4. Specifically, we need to clarify its practical implementation differences from Alt 1 and establish clear metrics for evaluation