AI Beam Model Reporting for Low-Overhead Downlink Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing NR downlink beam management methods require the transmission of CSI-RS or SSB signals on all Tx beams, consuming significant resources and imposing high measurement overhead and complexity on user equipment (UE).
Innovation Solution
Implementing artificial intelligence (AI) or machine learning (ML) models for downlink beam prediction, where user equipment (UE) reports model information to a network device, including details such as model identification, application scenario, and configuration, enabling efficient beam management and prediction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If CSI-RS or SSB signals are transmitted on all Tx beams, then downlink beam management can be performed, but resource consumption increases significantly
Solution Approach 1:
The patent extracts only the necessary reference signals for beam management rather than transmitting all possible Tx beams. The network device determines a first quantity of reference signals based on a second quantity of Tx beams, selecting only the essential subset needed for effective beam management, thereby reducing resource consumption while maintaining management capability
Solution Approach 2:
Instead of performing complete measurement on all Tx beams, the patent applies partial action by having the UE measure only a determined subset of reference signals. This partial measurement approach provides sufficient beam management information without the excessive resource cost of full beam sweeping
2Measurement precision
If CSI-RS or SSB signals are transmitted on all Tx beams, then beam measurement can be performed, but measurement overhead increases
Solution Approach 1:
The patent extracts only the necessary reference signals for beam measurement rather than requiring measurement of all Tx beams. The network device determines a first quantity of reference signals that is less than or equal to the second quantity of Tx beams, reducing the measurement overhead while maintaining sufficient measurement precision for beam selection
3Measurement precision
If CSI-RS or SSB signals are transmitted on all Tx beams, then beam measurement can be performed, but implementation complexity increases
Solution Approach 1:
The patent extracts and processes only a determined subset of reference signals rather than requiring the UE to process all Tx beams. This reduces the computational burden and implementation complexity at the UE side while maintaining the network device's ability to perform accurate beam measurement through intelligent selection of which reference signals to transmit and measure
Data Source
AI summary
The present disclosure provides a model information reporting method, device and apparatus, and a storage medium. The method comprises: a terminal determining first information, the first information being used for indicating one or more artificial intelligence or machine learning (AI/ML) models, and the AI/ML models being used for downlink beam prediction; and sending the first information to a network device.