RAN AI QoE Prediction Using UE Assistance Information
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing wireless communication systems lack the use of quality of experience (QoE) information and user equipment (UE) assistance information for artificial intelligence (AI) functions, leading to inefficiencies in predicting user equipment performance.
Innovation Solution
Implementing an AI model in network nodes to utilize QoE information and UE assistance information for model training and inference, including QoE measurement results, UE identifiers, and other relevant data to predict future QoE results.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If QoE information and UE assistance information are not used for AI functions, then the system operates with simpler processing, but the prediction accuracy of user equipment performance deteriorates
Solution Approach 1:
The system performs preliminary collection and preprocessing of QoE information and UE assistance information before AI model inference. Network nodes collect QoE measurements from UEs and assistive information from multiple sources (neighboring gNBs, core network) in advance, preparing the data infrastructure needed for accurate predictions without adding complexity during the actual prediction operation
Solution Approach 2:
The patent introduces an intermediary AI model layer between raw QoE/assistance information and network decision-making. This intermediary neural network model processes the complex input data (QoE measurements, UE assistance information, neighboring gNB data) and transforms it into actionable predictions, shielding the rest of the system from direct complexity while enabling accurate performance prediction
2Object-affected harmful factors
If QoE measurement information is collected from UE side, then the impact on Uu interface is minimized, but the complexity of data collection and processing increases
Solution Approach 1:
The patent extracts QoE measurement functionality from the network side and places it on the UE side. UEs autonomously perform QoE measurements and report results to network nodes, removing the measurement burden from the network infrastructure and minimizing impact on the Uu interface while distributing the measurement complexity to end devices that already possess the necessary capabilities
Solution Approach 2:
UEs are empowered to self-measure QoE parameters and self-report results to the network. The user equipment autonomously collects its own performance data, performs local processing, and provides reports without requiring complex network-side measurement infrastructure, thereby minimizing Uu interface impact while enabling comprehensive data collection
3Reliability
If AI model training and inference are performed using multiple information sources (QoE data, assistance information, neighboring gNB data), then prediction reliability improves, but the data processing load increases
Solution Approach 1:
The patent segments the AI processing function into distributed components across multiple network nodes (serving gNB, neighboring gNBs, core network). Each node collects and processes relevant local data independently, then shares results through standardized interfaces. This segmentation distributes the computational load while aggregating diverse data sources to improve prediction reliability
Solution Approach 2:
The patent creates a universal AI inference framework that can process multiple types of input data (QoE measurements, UE assistance information, neighboring gNB data, core network information) through a single neural network model. This multi-functional approach consolidates diverse data processing requirements into one unified system, improving reliability through comprehensive data utilization while avoiding the need for separate processing pipelines for each data type
Data Source
AI summary
Presented are systems and methods for predicting quality of experience of user equipment (UEs) using an artificial intelligence (AI) model. A first network node of a radio access network (RAN) may receive first assistance information for use with a first quality of experience (QoE) information to perform a first function of a neural network model from a second network node of the RAN. The first network node of the RAN may perform the first function using the first QoE information and the first assistance information.