R1-2410257
discussion
Discussion on specification support for beam management
From ETRI
Summary
ETRI's contribution presents 20 proposals and 3 observations on AI/ML-based beam management for NR air interface, covering configuration procedures, performance monitoring, measurement reporting, and consistency mechanisms for both UE-sided and NW-sided models across spatial and temporal domain beam prediction.
Position
ETRI advocates for a comprehensive framework supporting both UE-sided and NW-sided AI/ML models with emphasis on overhead reduction and consistency mechanisms. They strongly push FOR Options 1 and 2 over Option 3 for applicability procedures, L1-RSRP difference as a key performance metric, flexible Associated ID configurations across multiple locations, and differential RSRP reporting methods. They are pushing AGAINST overly complex signaling by preferring direct activation methods and supporting omission of RSRP values when differences are minimal to reduce overhead.
Key proposals
- Proposal 1 (Sec 4.1): For applicability for inference for the UE-side model, prioritize Option 1 and Option 2
- Proposal 2 (Sec 4.1): For the applicability functionality for UE-side model, the NW configures performance monitoring related parameters to the UE in Step 3
- Proposal 3 (Sec 4.2): For functionality activation of the UE-side model, the functionality can be activated directly if the CSI-reportConfig is configured in Step 5
- Proposal 5 (Sec 5.1): For UE-assisted performance monitoring, support the L1-RSRP difference between predicted and measured RSRP (Alt 3) as a performance metric
- Proposal 6 (Sec 5.2): For UE-assisted performance monitoring, support event reporting when counter exceeds threshold based on Top K beam prediction differences or L1-RSRP difference exceeding predefined value
- Proposal 8 (Sec 5.4): Support Monitoring IDs to differentiate monitoring processes for multiple UE-sided models
- Proposal 9 (Sec 6.1): For BM-Case2, support the configuration of multiple Resource Sets for Set A
- Proposal 11 (Sec 8.1): For the NW-sided model, reducing measurement overhead is necessary for model inference, especially in the case of temporal domain prediction
- Proposal 13 (Sec 8.2): Support methods for reducing UE reporting overhead during data collection for training when the AI/ML model is located on the NW-side through limiting maximum L1-RSRP values or specifying L1-RSRP threshold
- Proposal 14 (Sec 9.1): For the UE-side model, support the CSI report format for temporal domain beam prediction, including the optimal K beam information along with RSRP information from multiple time instances
- Proposal 16 (Sec 9.2): Support the indication field for model inference when an AI/ML model is located on the UE-side
- Proposal 18 (Sec 11.1): For the UE-sided model, the UE assumes that the same Associated ID represents a beam with the same direction
- Proposal 19 (Sec 11.2): For the UE-side model, the Associated ID can be configured in CSI-ReportConfig, CSI-ResourceConfig, or CSI-ResourceSet
- Proposal 20 (Sec 11.3): Support categorizing Associated IDs into cell-specific areas and areas that can be used across multiple cells