Joint Predictive Handover and Resource Allocation in Satellite-Terrestrial Integrated Networks

Satellite-terrestrial integrated networks (STINs) have emerged as a promising solution for providing global coverage and seamless connectivity. However, the high mobility of low Earth orbit (LEO) satellites creates a highly dynamic topology, posing severe challenges for mobility management and resource allocation. In such scenarios, inevitable delays between network measurement and decision execution often render STIN mobility management ineffective, which further causes outdated handovers, link failures, and degraded quality of service (QoS). To address these issues, we propose a joint predictive handover and resource allocation method based on the Long Short-Term Memory (LSTM)-Proximal Policy Optimization (PPO) algorithm. Specifically, by integrating LSTM into the PPO agent, the proposed method captures temporal dependencies in network dynamics and enables foresighted decisions that mitigate the negative effects of high mobility of LEO satellites and also reduce unnecessary handovers. Numerical simulation results show that compared to the benchmark methods, the proposed method achieves higher system sum rate, better load balancing, and lower link failure rates, implying that it enables more effective handover and resource allocation decisions through optimal service node selection and dynamic subcarrier assignment.