Selective CQI Learning for URLLC Reliability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current CQI reporting and link adaptation methods are insufficient for ensuring reliable and efficient communication, particularly in URLLC and AR/VR applications, as they do not adequately account for individual UE performance and can cause network overload and interference during exploration phases for learning CQI offsets.
Innovation Solution
Implementing a selective learning framework where the gNB controls exploration permission for specific UEs, allowing them to vary CQI reporting within configured boundaries to learn optimal offsets, while maintaining network performance and reliability, using machine learning techniques to adjust CQI reports based on SINR and interference patterns.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If all UEs perform CQI exploration and learning simultaneously, then learning accuracy and individual UE performance improvement are enhanced, but network overload and interference increase significantly
Solution Approach 1:
The patent applies local quality by enabling CQI exploration only for specific UEs that benefit most from it, rather than uniformly for all UEs. The gNB identifies and selects particular UEs based on their individual performance characteristics and channel conditions, allowing exploration only where it provides meaningful improvement. This selective approach maintains learning accuracy for targeted UEs while preventing network-wide overload and interference.
2Productivity
If CQI exploration is performed continuously, then optimal CQI offsets are learned more quickly, but transmission reliability and packet delay performance deteriorate
Solution Approach 1:
The patent implements periodic action by conducting CQI exploration in discrete, time-limited phases rather than continuously. The gNB configures exploration parameters including window lengths and periodically activates/deactivates exploration based on UE performance needs. This allows the system to alternate between exploration phases (for learning) and normal operation phases (for reliable transmission), achieving both learning speed and transmission reliability.
Solution Approach 2:
The patent applies dynamics by making the CQI exploration process adaptive and controllable. The gNB dynamically adjusts exploration activation, window lengths, and parameter configurations based on real-time network conditions and UE performance. This dynamic control allows the system to accelerate learning when conditions permit while maintaining transmission reliability when exploration would be harmful.
3Device complexity
If standard CQI reporting methods are used, then network compatibility and simplicity are maintained, but spectral efficiency and reliability for URLLC/AR-VR applications are insufficient
Solution Approach 1:
The patent applies self-service by enabling UEs to automatically learn and adapt their own CQI reporting characteristics through machine learning. The UE's ML model learns optimal CQI offsets based on individual channel conditions and performance feedback, allowing each UE to self-optimize its reporting without complex network configuration. This maintains relative simplicity while significantly improving spectral efficiency for URLLC and AR/VR applications.
Data Source
AI summary
A method includes transmitting an indication of support of learning reporting information; receiving a configuration related to the learning of the reporting information, the configuration comprising at least one parameter related to exploration of the reporting information; initiating the exploration of the reporting information to activate the learning of the reporting information; varying at least one value of the reporting information, based on the at least one parameter, to learn the at least one value of the reporting information; and reporting the at least one value of the reporting information during the learning or following a deactivation of the learning of the at least one value of the reporting information.