CSI Prediction Model Lifecycle Control in Wireless Terminals
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing wireless communication systems face challenges in efficiently managing artificial intelligence/machine learning (AI/ML) models, particularly in efficiently performing wireless signal transmission/reception procedures.
Innovation Solution
Implementing AI/ML models for channel state information (CSI) prediction and life cycle management (LCM) to enhance wireless signal transmission and reception, including configuring AI/ML models, monitoring their performance, and performing LCM procedures based on acquired information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If AI/ML models are used for CSI prediction to improve wireless signal transmission efficiency, then communication performance is improved, but model management complexity increases
Solution Approach 1:
The system implements feedback mechanisms where the terminal monitors AI/ML model performance by measuring CSI prediction accuracy, CQI variations, and SINR metrics. Based on this feedback, the terminal autonomously triggers LCM procedures such as model updates, switching, or fallback operations, creating a closed-loop system that improves communication efficiency while managing complexity through automated decision-making
Solution Approach 2:
The terminal performs self-service by autonomously monitoring its own AI/ML model performance and initiating LCM procedures without requiring constant network intervention. The terminal can request datasets for model updates, switch between models, or fallback to non-AI/ML operations based on its own performance assessments, reducing the signaling overhead and management complexity
2Reliability
If AI/ML model monitoring and LCM procedures are implemented, then model performance is optimized, but signaling overhead increases
Solution Approach 1:
The system implements partial monitoring by selectively measuring only the most critical performance metrics (CSI prediction accuracy, CQI variations, SINR) rather than comprehensive model analysis. LCM procedures are triggered only when performance thresholds are breached, performing just enough action to maintain reliability while minimizing signaling overhead
3Reliability
If frequent AI/ML model updates are performed to maintain prediction accuracy, then communication reliability is improved, but system resources are consumed
Solution Approach 1:
The system implements periodic LCM procedures triggered by performance monitoring rather than continuous updates. The terminal monitors model performance over measurement periods and initiates updates only when degradation exceeds thresholds, creating a periodic rather than continuous update cycle that maintains accuracy while conserving energy and computational resources
Data Source
Figure 1~2
Figure 3
Figure 4~5
AI summary
A terminal according to at least one of embodiments disclosed in the present specification may: configure an artificial intelligence/machine learning (AI/ML) model; obtain information about channel state information (CSI) prediction performance of the AI/ML model through monitoring of the AI/ML model; and on the basis of the obtained information about the CSI prediction performance, perform a life cycle management (LCM)-related procedure for the AI/ML model, wherein the LCM-related procedure may include at least one of: (i) transmitting information requesting a data set for updating the AI/ML model; (ii) transmitting information requesting configuration of a time interval in which the update of the AI/ML model is to be performed; (iii) transmitting information requesting switching of the AI/ML model; and (iv) transmitting information requesting a fallback using a non-Al/ML-based operation.