R1-2410255
discussion
Specification support for beam management
From Fraunhofer
Summary
This Fraunhofer document presents 17 proposals for AI/ML-based beam management in 5G NR, covering both UE-sided and network-sided models with focus on performance monitoring, configuration optimization, and overhead reduction. The proposals address two main use cases (BM-Case1 and BM-Case2) with comprehensive solutions spanning monitoring phases, beam configuration, measurement reporting, and quantization enhancements.
Position
Fraunhofer strongly advocates FOR UE-sided AI/ML models over network-sided models, emphasizing their advantages in reduced signaling overhead, shorter measurement-to-inference delay, and ability to use more measurements. They push FOR bitmap-based indexing as the most efficient beam reporting method and advocate FOR 2-phase monitoring with adaptive frequency. They are AGAINST beam indication for future time instances, arguing it offers little advantage while incurring high specification workload and considerable UE behavior modifications.
Key proposals
- Proposal 1 (Sec 2.1): For monitoring UE-sided models, support 2-phase monitoring with varying frequencies and reporting detail
- Proposal 5 (Sec 2.2): For BM-Case2, for inference, the reference time of the predicted time instances shall be determined based on the CSI reference resource or the transmission occasion of Set B resources
- Proposal 7 (Sec 2.3): Explore the possibility of using the resources configured for radio link monitoring for model monitoring purposes with respect to CSI beam reporting
- Proposal 9 (Sec 2.4): Study the support for radio link monitoring and link recovery procedures considering spatial and temporal beam prediction at the UE-side
- Proposal 10 (Sec 2.5): For the configuration of set A and set B of beams for UE-side BM-Case1, the relationship in use-cases where set B of beams is a subset of set A of beams shall be exploited for overhead reduction
- Proposal 12 (Sec 2.6): Support L3 measurements as a container for L1-RSRPs reporting for training/re-training purposes given its increased payload size, relaxed latency requirement and higher reliability
- Proposal 13 (Sec 2.7): For NW-sided models, adopt bitmap-based indexing for reporting measurements
- Proposal 14 (Sec 2.8): Consider multi-resolution quantization, increased step sizes, and adaptive reference beam for differential RSRPs quantization enhancement
- Proposal 15 (Sec 2.9): The use of a predicted beam that is not measured/received by the UE for beam indication is supported
- Proposal 16 (Sec 2.9): Beam indication for one or more future time instances is not supported
- Proposal 17 (Sec 2.10): For UE-sided models, for inference, study the UE reporting its inference time to the gNB
- Proposal 3 (Sec 2.1): Consider indication-based and event-based switching into a validation phase with events defined based on agreed performance metrics
- Proposal 6 (Sec 2.2): For BM-Case2, for UE-sided models, introduce a configuration that allows the exclusion of certain past measurements for the inference
- Proposal 8 (Sec 2.3): Support a model monitoring configuration that allows for collecting data for model training and monitoring of inactive models
- Proposal 11 (Sec 2.6): Prior to the selection of a container for data collection, study the AI/ML purposes separately in terms of amount of data, acceptable latency, security and reliability