R1-2410844
other
Session notes for 9.1 (AI/ML for NR Air Interface)
From CMCC
Summary
This is a session notes document from CMCC serving as Ad-hoc Chair for RAN1 #119 meeting on AI/ML for NR Air Interface. The document contains agreements and conclusions across 4 main areas: beam management, positioning accuracy enhancement, CSI prediction, and additional AI/ML studies including CSI compression and model/data aspects.
Position
CMCC as Ad-hoc Chair is facilitating consensus building across all AI/ML use cases for NR air interface. They are advocating FOR systematic specification support across beam management, positioning, and CSI applications while promoting standardized approaches for model structures, data collection, and inter-vendor collaboration. They are pushing FOR comprehensive liaison statements to other working groups to ensure feasibility of proposed solutions and coordinated development across 3GPP.
Key proposals
- Agreement (Sec 2): For UE-sided model, at least for BM-Case 1, the beam information in inference result report is CRI/SSBRI of resource in Set A
- Agreement (Sec 2): For both BM-Case 1 and BM-Case 2, for UE-sided model for inference, when Set A and Set B are configured within CSI report configuration, two CSI-ResourceConfigId s are configured for Set A and Set B separately
- Agreement (Sec 2): In Step 3, following configurations are provided from NW to UE: UE is allowed to do UAI reporting via OtherConfig, and the applicability report is based on CSI-ReportConfig for inference configuration and/or sets of inference related parameters
- Agreement (Sec 2): At least for monitoring Type 1 Option 2 of UE-side model monitoring, support to reuse CSI framework for the configuration for monitoring result report in L1 signaling
- Agreement (Sec 3): For the definition of sample-based measurement, for gNB/TRP measurement of an estimated channel response between a pair of UE and TRP, the starting time of the list of Nt consecutive samples is determined as starting time = first detected path rounded down with timing granularity T
- Agreement (Sec 3): For model performance monitoring of AI/ML positioning Case 1, support at least Option A where the target UE side performs monitoring metric calculation and may signal the monitoring outcome to the LMF
- Agreement (Sec 3): For Rel-19 AI/ML based positioning Case 3b, support enhancement to measurement with Nt' values of estimated channel response in time domain selected from Nt consecutive values with timing granularity T=2kxTc
- Agreement (Sec 3): For AI/ML based positioning Case 1, all assistance information from legacy UE-based DL-TDOA except info #7 can be provided from LMF to UE, with study of four alternatives for info #7 handling
- Conclusion (Sec 5.1): For Direction A 4-1 and 3a-1, send LS to RAN2 for the study of feasibility of parameter/dataset exchange, additionally CC SA2, SA3, SA5
- Agreement (Sec 5.1): For Direction A Option 3a-1 and Direction C, study the feasibility of scalable model structure specification over numbers of Tx ports, CSI feedback payload sizes, bandwidths, and number of slots
- Agreement (Sec 5.1): For studying the standardized model structure, adopt Transformer as backbone structure for temporal domain Case 0 with spatial-frequency domain input
- Agreement (Sec 5.1): For NW to collect data for training, study spec impacts including data format as codebook-based Rel-16 eType2 or Rel-18 eType2 for PMI prediction
- Agreement (Sec 5.2): For study of MI-Option2 (model identification with dataset transfer) for two-sided model, ID-X can be used for pairing the UE-part and the NW-part of a two-sided model
- Agreement (Sec 5.2): Regarding relationship of model ID, first indication, and second indication for model transfer/delivery Case z4, further study three options for their relationship
- Conclusion (Sec 5.2): For model delivery/transfer Case z4, if used directly for inference, study two candidate solutions to determine readiness of AI model with transferred parameters