R1-2409994
discussion
Discussion on AI/ML for beam management
From China Telecom
China Telecom's prior position on
9.1.1
at
RAN1#118bis
· AI-synthesized, paraphrased
verify sources →
Advocates for a comprehensive AI/ML beam management framework supporting both network-sided and UE-sided models with flexible data collection mechanisms and beam prediction accuracy KPIs as the primary performance monitoring metric.
Summary
China Telecom's contribution discusses AI/ML for beam management in NR, focusing on Life Cycle Management (LCM) aspects including data collection, model inference, and performance monitoring for both network-sided and UE-sided models. The document contains 16 technical proposals covering BM-Case 1 (spatial beam prediction) and BM-Case 2 (temporal beam prediction).
Position
China Telecom advocates FOR comprehensive support of multiple data collection options for flexibility across different LCM purposes (training, monitoring, inference), beam prediction accuracy KPIs as primary performance monitoring metrics, and full Set A measurement for accurate UE-assisted performance monitoring. They push AGAINST using only single resource set configurations for Set B in UE-sided models and deprioritize link quality metrics for AI/ML model performance monitoring, arguing these don't directly reflect prediction accuracy.
Key proposals
- Proposal 1 (Sec 2.1.1.1): Support M1 L1-RSRPs with beam indication, M2 L1-RSRPs, and M3 beam indices as contents of collected data for NW-sided AI/ML model
- Proposal 3 (Sec 2.1.2): For UE-side model, data collection can be triggered by the UE or initiated by network
- Proposal 4 (Sec 2.2.1.1): For NW-sided model inference, report content for beam related information can be L1-RSRP or L1-RSRP and beam ID
- Proposal 7 (Sec 2.2.2.1): For UE-sided model at least for BM Case-1, not support configuring only one resource set for Set B for inference results report
- Proposal 8 (Sec 2.2.2.2): For UE-sided model BM-Case1, support Option 2 for inference results report content - beam information on predicted Top K beams and RSRP of predicted Top K beams with highest predicted RSRP above proper probability
- Proposal 10 (Sec 2.2.2.3): For BM-Case 2 UE-side model, support Option 3 for reference time - based on latest transmission occasion of CSI-RS/SSB resource in Set B
- Proposal 11 (Sec 2.3.1.1): At least support beam prediction accuracy related KPIs as the metric for performance monitoring
- Proposal 12 (Sec 2.3.1.2): For UE-sided AI/ML model UE-assisted performance monitoring Alt 1, configure resource set for measurement as the full Set A
- Proposal 14 (Sec 2.3.1.3): If number of AI/ML model failures reaches predefined threshold, consider performing model switch or fallback to non-AI/ML method
- Proposal 16 (Sec 2.4): For associated ID configured in CSI framework for UE-sided model, associated ID is configured per CSI report configuration