R1-2410217
discussion
Support for AI/ML for positioning accuracy enhancement
From Sony
Summary
Sony's contribution presents a comprehensive framework for AI/ML-enhanced positioning in NR, covering the entire AI/ML lifecycle from data collection to model deployment and monitoring. The document contains 15 detailed proposals addressing key aspects including CIR-based data collection, model transfer mechanisms, consistency between training and inference, and performance monitoring across different AI/ML positioning cases.
Position
Sony strongly advocates for sample-based time domain channel measurements over path-based measurements for AI/ML model input, emphasizing information richness for positioning accuracy. They push for comprehensive CIR reporting mechanisms and advocate for robust consistency mechanisms between training and inference phases, particularly supporting centralized model training at LMF with distributed inference at UE/gNB. Sony favors Option A approaches for performance monitoring where UE/gNB perform local model validity assessment.
Key proposals
- Proposal 1 (Data Collection): Support radio channel characteristics reporting in a form of channel impulse response (CIR) for AI/ML positioning
- Proposal 2 (Data Collection): Support configurable CIR measurement report (e.g., report size, measurement window size) to reduce signalling overhead
- Proposal 4 (Data Collection): Support model transfer operation from server node (e.g., LMF) to UE/gNB to support AI/ML-assisted positioning
- Proposal 6 (Model Input): Support sample-based time domain channel measurements as the AI/ML model input
- Proposal 7 (Model Output): For AI/ML assisted positioning case 2a and case 3a, positioning measurement reporting includes LOS/NLOS indicator using legacy reporting format with indication whether measurement is based on AI/ML computation
- Proposal 8 (Model Inference): Define a set of parameters (e.g., part/all of DL-PRS configuration, received signal quality) representing reference signal characteristics for AI/ML positioning
- Proposal 9 (Model Inference): Support association of reference signal characteristics to trained AI/ML model and inference operation to ensure consistency
- Proposal 11 (Model Inference): Assistance data information elements for maintaining consistency between model training and inference are at least those shown in Table 1
- Proposal 13 (Performance Monitoring): For case 1, support Option A where LMF sends criteria (parameters for performance evaluation, thresholds) to UE performing performance metrics computation
- Proposal 14 (Performance Monitoring): For case 3a with option B, support LMF providing indication of AI/ML model validity to NG-RAN node
- Proposal 15 (Performance Monitoring): For case 1 and case 3a with option A, UE or gNB to provide indication of AI/ML model validity to AI/ML server/management