R1-2410149
discussion
AI/ML based Beam Management
From Google
Summary
This 3GPP RAN1 technical document from Google presents 27 proposals for AI/ML-based beam management enhancements in 5G NR, covering beam measurement, reporting, indication, failure recovery, and performance monitoring across both network-side and UE-side AI/ML models.
Position
Google advocates FOR comprehensive AI/ML beam management support including both network-side and UE-side models, pushing for flexible reference signal configurations (AP-CSI-RS, SSB), enhanced reporting mechanisms with confidence indicators, separate TCI state pools for AI/ML predictions, and practical implementation considerations like CPU resource management and performance monitoring relaxation. They strongly support Option 1 for inference configuration management to avoid multi-stage RRC configuration latency issues.
Key proposals
- Proposal 1 (Sec 2): Support to configure AP-CSI-RS for BM as RS for Set B for BM-Case 2
- Proposal 2 (Sec 2): Support to configure SSB as RS for Set B for both BM-Case 1 and BM-Case 2
- Proposal 3 (Sec 3.1): Support to report a confidence level indicator for L1-RSRP report to indicate the maximum L1-RSRP measurement error for each beam
- Proposal 6 (Sec 3.2): For beam report based on UE model inference for SD beam prediction, support top-K beam selection up to UE implementation, predicted RSRP based on reference transmission power, and both probability and confidence information reporting
- Proposal 7 (Sec 3.2): An inference report for BM-Case 1 should take 1 eCPU and 1 legacy CPU, with eCPU for inference and legacy CPU for Set B RS measurement
- Proposal 8 (Sec 4.1): Support dynamic activation/deactivation of periodic TRS with regard to TCI activation/indication based on the predicted beam
- Proposal 10 (Sec 4.2): Support to configure separate TCI state pools for legacy beam indication and TCI state for beam prediction
- Proposal 13 (Sec 4.2): Support UE feedback before the beam action time for performance validation for predicted beam in addition to ACK/NACK for TCI update signaling
- Proposal 14 (Sec 5): For BFR based on NW side model for SD beam prediction, support UE initiated beam report with L1-RSRP for configured SSBs/CSI-RSs when UE cannot identify candidate beam
- Proposal 18 (Sec 6): Support to configure aperiodic CSI-RS as set A RS, triggered by group-cast DCI
- Proposal 20 (Sec 7): For UE-side model monitoring, support periodic beam prediction accuracy evaluation with counter-based reporting using SCell BFR mechanism
- Proposal 23 (Sec 7): Support performance monitoring relaxation for UE in low mobility and high beam quality state
- Proposal 25 (Sec 8): Support to configure associated ID per CSI report configuration with dynamic activation/deactivation
- Proposal 26 (Sec 9): Support ML based beam prediction to facilitate event triggered beam report
- Proposal 27 (Sec 10): Support Option 1 for applicability for inference for UE-side model