R1-2500275
discussion
Discussion on specification support for positioning accuracy enhancement
From CMCC
CMCC's prior position on
9.1.2
at
RAN1#119
· AI-synthesized, paraphrased
verify sources →
Slightly prefers sample-based measurements to avoid potential errors from intermediate path-based processing at the UE, proposing that the entity deriving the AI model should provide recommended sample numbers. Argues that offline training with marginal specification impact is feasible but questions the feasibility of obtaining ground-truth labels via PRUs due to large dataset sizes. Proposes reinterpreting existing Timing Measurement Quality IEs for gNB-side models and reusing legacy LOS/NLOS reporting formats, while deprioritizing certain consistency options in favor of further study.
Summary
CMCC discusses specification impacts for AI/ML-based positioning in NR, focusing on measurement types, data collection, and model monitoring. The document contains 20 proposals and 13 observations covering sample-based vs. path-based measurements, ground-truth label acquisition, and consistency between training and inference.
Position
CMCC slightly prefers sample-based measurements for Case 2b, arguing that they are closer to UE implementation and avoid errors from intermediate path-based processing. They propose supporting candidate values for Nt' such as [9 16 24] or any value in [1 24], and require k to support at least (0…5) to avoid complex interpolation algorithms. For Case 2a, they propose that reported timing information should be based on the legacy reporting format to minimize LMF modifications. They suggest reinterpreting the existing 'Timing Measurement Quality' IE for gNB-side model performance metrics and using legacy LOS/NLOS reporting for AI-assisted positioning. Regarding model consistency, they propose further discussing Options 3 and 4 while deprioritizing Options 1 and 2, and emphasize the need to clarify the granularity of validity areas and RS configuration consistency.
Key proposals
- Proposal 1 (Methodology/Considerations): Sample-based measurements for AI/ML positioning is slightly preferred for case 2b.
- Proposal 2 (Considerations): For candidate value of Nt', whether less candidate value like [9 16 24] or any value between [ 1 24 ] could be considered.
- Proposal 3 (Considerations): For candidate value of k, at least (0…5) should be supported.
- Proposal 4 (Considerations): Separately reporting transmit offset of the first detected path besides Nt' may not necessary.
- Proposal 5 (Considerations): For path-based measurement reporting, current DL PRS-RSRPP and UL-SRS-RSRPP can be used as a starting point.
- Proposal 6 (Considerations): For AI/ML based positioning, it needs more discussion on the feasibility of obtaining the ground-truth label via PRUs, in which case the training dataset size is large.
- Proposal 7 (Considerations): For AI/ML based positioning, more discussion is needed for the comparison between CIR and PDP as model inputs.
- Proposal 10 (Considerations): For AI/ML based positioning, whether the reported measurement is AI based could have an indication.
- Proposal 11 (Considerations): For AI/ML assisted positioning, existing IE “Timing Measurement Quality” can be reinterpreted for reporting performance metric for gNB side model.
- Proposal 15 (Considerations): For AI/ML based positioning, the relationship between model monitoring and positioning integrity can be considered.
- Proposal 16 (Considerations): For AI/ML based positioning, further discuss option 3 and 4. Deprioritize option 1 and 2.
- Proposal 18 (Additional spec impact): For AI/ML assisted positioning case 2a, the reported timing information should be based on the legacy reporting format.
- Proposal 19 (Additional spec impact): For AI/ML based positioning case 2b, a reference time for determining the measurement starting point is needed.
- Proposal 20 (Additional spec impact): For AI/ML based positioning case 2b, definition of sample-based measurement reporting for case 3b could be used as a starting point.