FL-SDMN: Federated Learning-Based Software-Defined Mobile Networking for Intelligent Handover Optimization

The rapid growth in mobile users and the increasing adoption of data-intensive applications have created high network demands for low-latency and reliable mobility management in mobile communication networks. However, the traditional handover approaches are rule-based and rely solely on signal strength thresholds with hysteresis margins, which are prone to ping-pong effects and limiting their ability to adapt to dynamic network conditions. Machine Learning (ML) models have been integrated for handover prediction, but their centralized architecture compromises user data privacy, potentially conflicting with the General Data Protection Regulation (GDPR). The centralized ML approaches also introduce scalability constraints that limit their effectiveness in dense network deployments. To address these challenges, this work proposes a Federated Learning framework for Software-Defined Mobile Networking (FL-SDMN) that combines privacy-preserving federated learning with centralized network coordination to enable intelligent handover optimization in 5G and beyond networks. The framework leverages a lightweight federated ExtraTrees ensemble model with weighted tree-based aggregation and SDMN controller to provide global network coordination. The federated learning framework has a three-layer decision pipeline that transforms handover control from a reactive threshold-based mechanism into a predictive, A3-derived optimization process without TTT persistence. The model was trained using real-world 5G mobility data, while handover performance, controller decision latency and scalability were evaluated separately using a discrete-time system level simulator. The findings demonstrate that integrating FL, Extra Trees, and SDMN provides a scalable and deployment-oriented approach to intelligent mobility management for 5G and beyond networks.