R1-2409957
discussion
Discussion on specification support for beam management
From Panasonic
Summary
Panasonic's technical document for 3GPP RAN1 discusses AI/ML-based beam management for NR air interface, presenting 12 detailed proposals covering UE-side and NW-side model inference, consistency mechanisms, and performance monitoring. The document addresses technical specifications for both spatial-domain (BM-Case1) and temporal (BM-Case2) DL Tx beam prediction scenarios.
Position
Panasonic strongly advocates FOR Option 2 approach for UE-side model inference applicability (against Options 1 and 3), arguing it provides better optimization and efficiency by avoiding unnecessary CSI-ReportConfig IEs. They push FOR cross-cell applicability of 'associated IDs' (against cell-level restrictions) while proposing MNO-level management to preserve proprietary information. They advocate FOR flexible dedicated configurations for performance monitoring rather than embedded approaches, and FOR extending current beam reporting limitations beyond 4 beams to support AI/ML requirements.
Key proposals
- Proposal 1 (Sec 2.1): Support Option 2 for UE-side model inference applicability, using inference-related parameters based on CSI-ReportConfig structure but configured separately, including Set A/B information, report content, and time instances for BM-Case 2
- Proposal 3 (Sec 2.2.2.1): For NW-sided model inference, support configuring measurement windows with measurement resource sets to define measurement duration for aperiodic CSI triggers
- Proposal 4 (Sec 2.2.2.2): Enhance group-based beam reporting to support reporting more than 4 beams in one report for NW-sided models, extending current 4-beam limitation
- Proposal 5 (Sec 2.2.2.3): Extend Rel. 17 TCI state activation signaling to activate K predicted beams for N future time instances in BM-Case 2, with options for combined or separate TCI state sets
- Proposal 7 (Sec 2.2.3.3): Differentiate prediction and measurement results through either prediction-related metrics as reporting quantities or different resource sets for prediction versus measurement
- Proposal 8 (Sec 2.2.3.4): Support mapping/association of beams between Set A and Set B based on QCL relationship to improve UE-sided model inference performance
- Proposal 9 (Sec 2.3): Apply 'associated IDs' concept for multiple cells to ensure consistency of NW-side additional conditions across training and inference for UE-sided models in both BM-Case 1 and BM-Case 2
- Proposal 10 (Sec 2.3): Determine 'associated IDs' within a single NW operator (MNO) to preserve proprietary information while enabling cross-cell applicability
- Proposal 11 (Sec 2.4.1): Configure dedicated resource sets in dedicated CSI report configurations for Option 2 UE-assisted performance monitoring, separate from inference configurations
- Proposal 12 (Sec 2.4.2): Enhance group-based beam reporting to support performance monitoring for NW-sided models, enabling monitoring of more than 4 beams