Cell Handover QoE Prediction for Target Cell Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing handover methods in cellular networks, such as those based on fixed thresholds or general KPI forecasting, fail to provide accurate guidance, leading to degraded user experience and QoE for both the moving UE and existing users in the target cell.
Innovation Solution
Implement a machine learning model that predicts the quality of experience (QoE) for both a moving UE and existing users in neighbor cells, using multi-target regression to optimize handover decisions by evaluating cell and UE KPIs, and applying a cell scoring formula to select the best target cell.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If fixed threshold handover methods are used, then handover decisions are simple to implement, but handover effectiveness and user experience quality are degraded
Solution Approach 1:
The patent transforms handover decision-making from static threshold-based parameter comparison to dynamic machine learning model predictions. The system uses ML models to predict future QoE metrics (throughput, latency, packet loss) based on historical and real-time data, allowing adaptive parameter selection that optimizes both reliability and effectiveness while maintaining implementation feasibility through standardized ML pipelines.
Solution Approach 2:
The patent introduces machine learning models as intermediary components between raw network measurements and handover decisions. These ML models process multiple input features (signal strength, load, historical performance) and generate predicted QoE outcomes, serving as a intelligent mediator that bridges simple measurement collection and complex decision-making, thereby improving reliability without significantly increasing operational complexity.
2Adaptability or versatility
If general KPI forecasting is used, then prediction coverage is broad, but prediction accuracy for individual user experience is insufficient
Solution Approach 1:
The patent implements a multi-layered prediction approach where general network-level KPI forecasting provides broad coverage, while user-specific ML models deliver precise individual QoE predictions. The system maintains multiple prediction models operating at different granularities (network-wide trends, cell-level patterns, user-specific behavior) and combines them to achieve both comprehensive coverage and high accuracy for individual user experience assessment.
Solution Approach 2:
The patent segments the prediction system into multiple independent ML models that operate at different levels: network-level forecasting for overall trends, cell-level models for local conditions, and user-specific models for individual QoE prediction. This segmentation allows each model to specialize in its domain, improving overall prediction accuracy while maintaining broad coverage through the hierarchical structure.
3Speed
If handover is performed without predictive analysis, then handover speed is fast, but user experience quality and network stability are degraded
Solution Approach 1:
The patent performs preliminary QoE prediction and handover suitability assessment using machine learning models before actual handover execution. The system evaluates multiple candidate target cells, predicts future QoE outcomes for each, and pre-selects the optimal target before triggering the handover. This preliminary analysis ensures that fast handover execution is always directed toward cells that will maintain or improve user experience and network stability.
Solution Approach 2:
The patent implements a feedback mechanism where actual post-handover QoE measurements are compared against ML-predicted values, and this feedback is used to continuously refine and retrain the prediction models. The system monitors handover outcomes, identifies prediction errors, and uses this information to improve future predictions, thereby enhancing network stability and user experience while maintaining fast handover speeds through increasingly accurate predictive guidance.
Data Source
AI summary
A system can, for respective neighbor cells of neighbor cells of a cell that communicates with user equipment, use a trained machine learning model to predict respective first quality of experience values that the user equipment is predicted to receive while communicating with the respective neighbor cells, and respective second quality of experience values for respective existing user equipment in the respective neighbor cells in a case where the user equipment has communicated with the respective neighbor cells. The system can determine respective scores for the respective neighbor cells based on the respective first quality of experience values and the respective second quality of experience values. The system can perform a handover of the user equipment from the cell to a selected neighbor cell of the neighbor cells based on the selected neighbor cell being determined to have at least a threshold high score among the respective scores.