R1-2500391
report
AI/ML for beam management
From Ericsson
Ericsson's prior position on
9.1.1
at
RAN1#119
· AI-synthesized, paraphrased
verify sources →
Advocates for a comprehensive AI/ML framework that leverages existing CSI mechanisms with minimal specification impact while maximizing functionality. Strongly supports expanding aperiodic CSI-RS resources from 16 to 64 beams to enable practical AI/ML deployment. Proposes uncertainty quantification in UE predictions to enable trustworthy AI/ML and adaptive Top-K beam selection based on model confidence. Supports extensive overhead reduction mechanisms for NW-sided models. Opposes complex new signaling frameworks, favoring reuse of existing CSI infrastructure with targeted enhancements.
Summary
Ericsson presents 21 proposals and 6 observations for AI/ML beam management in NR, focusing on UE-sided model configuration, inference reporting, performance monitoring, and NW-sided data collection overhead reduction. The document argues for configuring associated IDs at the ResourceSet level rather than CSI-ReportConfig to support larger beam sets and flexible aperiodic triggering, while also proposing mechanisms for uncertainty reporting, adaptive Top-K selection, and joint activation of monitoring and inference configurations.
Position
Ericsson proposes configuring the associated ID at the CSI-ResourceSet level rather than the CSI-ReportConfig level to preserve fundamental CSI framework assumptions and enable predictions beyond current resourceSet size limits. They require support for separate CSI-ResourceConfigIds for Set A and Set B, and propose extending aperiodic resource sets to 64 NZP CSI-RS resources. For inference reporting, Ericsson supports adaptive Top-K values and the inclusion of probability and RSRP confidence intervals to handle model uncertainty and data drift. They argue for joint activation of monitoring and inference configurations to minimize signaling overhead and propose specific performance metrics based on Top-1/Top-K alignment. For NW-sided models, they propose mechanisms to reduce reporting overhead, including pre-processing Set B beams and omitting duplicated or unstable training data samples.
Key proposals
- Proposal 1 (Sec 2.1): Configure the associated ID separately for Set A and Set B in the corresponding ResourceSet to avoid limiting aperiodic configurations to a single resource set.
- Proposal 2 (Sec 2.2): Support Alternative 3 for Set A/B configuration, where two separate CSI-ResourceConfigIds are configured for Set A and Set B.
- Proposal 3 (Sec 2.3): Enable the network to specify a Set A beam subset restriction similar to Codebook Subset Restriction (CBSR) to manage inter-cell interference.
- Proposal 4 (Sec 2.4): Use CSI-ReportConfig for data collection with a new field 'resourcesForDataCollection' indicating Set A or Set B, allowing the UE to measure both sets without reporting.
- Proposal 5 (Sec 2.5): Extend aperiodic resource configuration to support a resourceSet with 64 NZP CSI-RS resources, addressing the current limit of 16.
- Proposal 6 (Sec 2.6): Support Option 1 for the reference time of inference results in BM-Case 2, based on the uplink slot for the report.
- Proposal 7 (Sec 3.1): Support UE indication of invalid or inaccurate inference results in the report, e.g., by setting CRI/SSB-RI to a fixed value.
- Proposal 8 (Sec 3.2): Support reporting probability-related information (Option 3) and RSRP confidence intervals (Option 4) to capture model uncertainty and data drift.
- Proposal 9 (Sec 3.3): Support an adaptive value of K for Top-K beam reporting, based on the UE-sided model output to ensure a certain probability threshold.
- Proposal 10 (Sec 3.4): Allow the UE to update reported inference results for future time instances in BM-Case 2 if predictions become invalid due to P2 procedures or channel changes.
- Proposal 11 (Sec 4.1): Support performance metric Option 2, where the Top-1 measured beam is one of the Top-K predicted beams.
- Proposal 13 (Sec 4.2): Support joint activation of the monitoring and linked inference report configuration to reduce signaling overhead.
- Proposal 16 (Sec 5.1): Consider multiple options for P2 sweep enhancements, including NW indicating beam IDs/TCI states for the P2 ResourceSet based on UE predictions.
- Proposal 19 (Sec 6.2): Support NW configuration for UEs to pre-process Set B beams by reporting only beams within X dB of the strongest or at most N strongest beams to reduce overhead.
- Proposal 21 (Sec 6.3): Study omission/selection of collected data for NW-sided models, such as avoiding signaling duplicated samples or data during large channel variations.