UE AI/ML Control for Low-Overhead CSI and Beam Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing mobile communication systems face challenges in efficiently utilizing AI/ML models for wireless communication, particularly in reducing overhead and power consumption while maintaining accurate channel state information feedback, beam management, and positioning accuracy.
Innovation Solution
Implementing AI/ML technology within user equipment (UE) to perform model training and inference, allowing for reduced reference signal transmission and optimized resource usage, such as CSI-RS and PRS, through training and inference modes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional reference signal transmission is used for channel state information feedback, then measurement precision is maintained, but overhead and power consumption increase
Solution Approach 1:
The patent uses AI/ML models to generate inferred channel state information that copies the functionality of traditional reference signal measurements. Instead of continuously transmitting reference signals, the UE generates CSI feedback by inferring from previously received reference signals and current channel conditions, maintaining accuracy while reducing overhead.
Solution Approach 2:
The patent implements partial reference signal transmission where reference signals are sent only when necessary (e.g., when channel conditions change significantly) rather than continuously. The AI/ML model compensates for the reduced measurements by inferring channel state between reference signal transmissions, achieving energy savings without sacrificing feedback accuracy.
2Loss of energy
If AI/ML models are implemented in user equipment for channel state information inference, then overhead is reduced, but device complexity increases
Solution Approach 1:
The patent divides the AI/ML processing functionality into separate modules within the UE: a model reception module that receives pre-trained models from the network, a model execution module that performs inference, and a feedback generation module that creates CSI reports. This segmentation allows the complex AI/ML functionality to be integrated incrementally without overwhelming the existing UE architecture.
Solution Approach 2:
The network performs model training in advance and delivers pre-trained AI/ML models to the UE. The UE then only needs to execute the pre-trained models for inference rather than performing computationally intensive training operations. This preliminary action by the network significantly reduces the processing complexity required at the UE side.
3Loss of energy
If reference signal transmission is reduced for power savings, then power consumption decreases, but measurement precision deteriorates
Solution Approach 1:
The patent implements a feedback mechanism where the UE sends inferred CSI reports to the network based on AI/ML model predictions. The network uses this feedback to update and retrain models, which are then redistributed to UEs. This closed-loop feedback ensures that even with reduced reference signal transmissions, the system maintains measurement precision through continuous model improvement based on actual channel conditions.
Data Source
AI summary
In an aspect, a communication control method is a communication control method in a mobile communication system. The communication control method includes a step of transmitting to a base station, by a user equipment, at least either of a use condition indicating a condition for using a plurality of respective AI/ML models or an execution condition indicating a condition for executing an operation for the plurality of respective AI/ML models.