AI/ML Mobility Prediction for Robust Wireless Handovers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional mobile voice communication networks face challenges in handling high UE mobility and frequent handovers, leading to issues such as handover failures, radio link failures, and increased power consumption due to reactive measurement schemes, which are inadequate for next-generation communication systems.
Innovation Solution
Implementing AI/ML assisted mobility in wireless communication systems by utilizing UE-side and network-side models for proactive prediction of cell and beam measurements, enabling improved handover decisions and reducing unintended events.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional reactive measurement schemes are used for handover decisions, then the system maintains simple measurement mechanisms, but handover failures and radio link failures increase due to inadequate response to high UE mobility
Solution Approach 1:
The patent implements AI/ML models that perform preliminary prediction of cell and beam measurements before handover events occur. The UE and network use trained models to forecast future measurement values and identify potential handover candidates in advance, enabling proactive handover decisions rather than reactive responses. This preliminary action reduces handover failures by preparing handover candidates before mobility events trigger traditional measurement schemes.
2Reliability
If frequent measurements are performed to handle high UE mobility, then handover decisions can be made more accurately, but power consumption increases significantly
Solution Approach 1:
The patent applies partial action by using AI/ML predictions to identify only the most promising handover candidates rather than performing exhaustive measurements on all possible cells. The system performs measurements selectively based on prediction outcomes, evaluating quality metrics only for predicted candidate cells. This partial measurement approach maintains handover decision accuracy while significantly reducing power consumption compared to frequent comprehensive measurements.
3Reliability
If AI/ML models are deployed for proactive measurement prediction, then handover robustness improves, but device and network complexity increases
Solution Approach 1:
The patent segments the AI/ML functionality into separate UE-side models and network-side models with distinct responsibilities. The UE model performs local prediction of cell measurements and identifies candidate cells, while the network model evaluates detailed quality metrics and makes final handover decisions. This segmentation distributes complexity across multiple components rather than concentrating it in a single system, making the overall system more manageable and deployable.
Solution Approach 2:
The patent introduces AI/ML prediction outputs as an intermediary layer between traditional measurement schemes and handover decision-making. The prediction models generate candidate cell lists and predicted measurement values that serve as intermediaries, filtering and prioritizing information before it reaches the handover decision logic. This intermediary approach simplifies the decision-making process by reducing the volume of data that requires detailed evaluation.
4Reliability
If traditional measurement schemes are used, then the system maintains simple operation procedures, but unintended handover events and handover failures increase
Solution Approach 1:
The patent implements feedback mechanisms where the UE reports predicted measurement outcomes and candidate cell information to the network, which then adjusts handover parameters and provides correction signals. The system continuously monitors handover performance and uses this feedback to refine AI/ML model predictions and tuning parameters. This feedback loop improves handover success rates by learning from past events while maintaining operational simplicity through automated adjustments rather than manual configuration changes.
Data Source
AI summary
Methods, systems, and apparatuses are provided for Artificial Intelligence/Machine Learning (AI/ML) assisted mobility in a wireless communication system, wherein a method for a User Equipment (UE) comprises receiving a first configuration of an AI/ML functionality for a measurement prediction, and performing at least one action based on at least an evaluation of quality of a cell from the measurement prediction, wherein the at least one action includes reporting an outcome of the evaluation to a network.