R1-2407746
discussion
Specification support for beam management
Summary
This 3GPP RAN1 technical document from Tejas Networks presents 33 proposals and 6 observations for AI/ML-based beam management in 5G NR, covering both UE-sided and network-sided models for spatial (BM-Case1) and temporal (BM-Case2) beam prediction. The document addresses key aspects including Associated ID configuration for training/inference consistency, performance monitoring mechanisms, reporting optimization, and TCI state management.
Position
Tejas Networks advocates FOR simplified beam management through single resource set configurations where possible (Set B only for UE-sided BM-Case1), differential L1-RSRP reporting to reduce overhead, and flexible Associated ID mechanisms supporting both CSI framework and higher-layer signaling. They push FOR dynamic configurability in prediction windows, quantization step sizes, and resource allocation while supporting standardized maximum values (M=256 beams). They advocate AGAINST overly complex dual resource set configurations when Set B alone is sufficient, and push for practical TCI state limits (N=2-4) balancing capability with signaling overhead.
Key proposals
- Proposal 1 (Associated ID Config): Associated ID should be configured as part of CSI-Report Config and mapped to resources in Set A (predicted beams) and Set B (measured beams)
- Proposal 6 (Performance Monitoring): For BM-Case1/2 with UE-side AI/ML model, L1 signaling can be used to send measurement results to NW for Type 1 performance monitoring
- Proposal 9 (NW-sided Reporting): For NW sided model, support multiple resource sets associated to one L1 beam report with Set A reporting beam IDs and Set B reporting L1-RSRP measurements
- Proposal 11 (Future Prediction): For UE-sided model BM-Case 2, support configuring UE with N future time instances for inference with reference time aligned to latest Set B measurement report
- Proposal 13 (Resource Set Simplification): For UE-sided model at least for BM case-1, support only one resource set (Set B) configured for CSI report generation and inference results reporting
- Proposal 15 (Differential Reporting): For NW-sided model BM-Case 2, support reporting Differential L1-RSRP for model inference to reduce overhead
- Proposal 19 (Maximum Beams): Consider M = 256 as starting point for maximum number of beams in NW-sided model inference reporting
- Proposal 21 (Quantization): For NW-sided model differential L1-RSRP reporting, support larger quantization step size to reduce overhead
- Proposal 25 (UE-sided Ranking): For UE-sided model BM-Case 1, ranking information of predicted Top K beams conveyed through order of reported information based on predicted RSRP or probability
- Proposal 30 (TCI States): For BM-Case 2, extend Rel-17 TCI state activation methods with N joint TCI states where N=2-4 is reasonable considering complexity and signaling overhead
- Proposal 32 (TCI Configuration): For TCI state configuration in BM-Case 2, use historic TCI state for predicting N future TCI states along with Set B historic measurements
- Proposal 16 (Dynamic Configuration): For NW-sided model BM-case 2, support dynamic configuration approach allowing gNB to request data from specific time instances most relevant for beam prediction