Adaptive Handover Optimization in LEO Satellite Networks Using Energy-Aware Q-Learning

Low Earth orbit (LEO) satellite networks are rapidly becoming a foundational infrastructure for global connectivity, especially in remote and underserved regions. However, the high mobility of LEO satellites results in frequent handovers, leading to increased signaling overhead, elevated energy consumption and potential service disruptions. To address these challenges, we propose an energy-aware handover management framework based on Q-learning, specifically tailored for single-user, remote scenarios. The proposed scheme dynamically adapts to real-time network conditions, aiming to maximize signal quality while minimizing unnecessary handovers and their energy cost. We evaluate its performance against three baseline methods: traditional Q-learning, predictive modeling, and entropy-based selection. Simulation results show that our approach improves cumulative signal quality by over 50%, reduces handover frequency by up to 65% and achieves higher coverage probability compared to baseline methods, demonstrating its effectiveness in enhancing both service continuity and energy efficiency in LEO satellite communications.