Wireless base stations and wireless communication methods
The wireless base station's AI/ML model unit addresses the expiration and performance degradation of AI/ML models by providing real-time feedback on handover and link failures, facilitating timely retraining and enhancing UE mobility performance.
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- NTT DOCOMO INC
- Filing Date
- 2026-04-09
- Publication Date
- 2026-07-07
AI Technical Summary
Existing AI/ML models in wireless communication systems have an expiration date and deteriorate in performance, necessitating retraining to maintain optimal mobility of user equipment (UE), but current methods lack efficient mechanisms for timely retraining.
A wireless base station equipped with an AI/ML model unit that monitors performance and transmits feedback on handover, radio link failures, and beam failures to an operations and maintenance entity, enabling dynamic retraining based on occurrence rates and performance thresholds.
This approach ensures optimal mobility of user equipment by maintaining the performance of AI/ML models, minimizing call drops, radio link failures, and unnecessary handovers through timely retraining.
Smart Images

Figure 2026113692000001_ABST