O-RAN Traffic Steering xApp for Per-UE Handover Optimization

The patent introduces an intermediary AI/ML-based handover optimization system that sits between traditional RRM heuristics and actual handover decisions. This intermediary layer processes network state information, predicts optimal handover targets, and generates recommendations that are then executed by the RAN controller, thereby improving optimization accuracy without requiring complete redesign of the existing RAN architecture.

The handover optimization function is segmented into distinct modular components: data collection module, AI/ML model module, decision generation module, and execution module. Each component performs a specific function and can be independently developed, trained, and deployed. This segmentation allows the complex optimization task to be broken down into manageable parts while maintaining high accuracy through the coordinated work of specialized modules.

The patent implements a universal AI/ML-based optimization framework that can handle multiple UE types (eMBB, URLLC, mMTC) and various network conditions through a single multi-functional system. The same core architecture processes handover decisions for different UEs, adapting to specific requirements through configurable parameters and trained models rather than requiring separate specialized systems for each UE type or scenario.

The system uses simulation and digital twin techniques to create virtual copies of the network environment for training and testing AI/ML models before deployment. By copying network states and behaviors in a virtual environment, the system can extensively monitor, measure, and optimize handover decisions without impacting real network performance, thereby reducing the actual monitoring complexity deployed in production.