现代电子技术2026，Vol.49Issue(11)：20-24,31,6.
基于深度强化学习多目标低轨卫星切换
Multi-objective low earth orbit satellite handover based on deep reinforcement learning
摘要
Abstract
The high-speed movement of low earth orbit(LEO)satellites causes users to handover satellites frequently,which increases system overhead and signaling load.Meanwhile,the dynamic network topology and high user mobility further complicate mobility management.In view of the above,this paper proposes an AI-based satellite handover optimization strategy that integrates minimized handover frequency,satellite response time,and maximized system throughput.The deep reinforcement learning is used to realize intelligent decision.The method constructs a multidimensional reward function combining signal-to-noise ratio(SNR),service duration and response time.In addition,the long short-term memory(LSTM)network is introduced to enhance the capabilities of temporal modeling.Simulations show that,in the case of different user scales and scenarios,the proposed method outperforms existing approaches in improving throughput and reducing handover frequency and failure rate,which demonstrates its effectiveness in ensuring communication continuity and system stability.关键词
低轨卫星/星间切换/多目标优化/切换算法/深度强化学习/长短期记忆神经网络Key words
LEO satellite/inter-satellite handover/multi-objective optimization/handover algorithm/deep reinforcement learning/LSTM neural network分类
信息技术与安全科学引用本文复制引用
宋美欣,田金凤,张汉中,周婷..基于深度强化学习多目标低轨卫星切换[J].现代电子技术,2026,49(11):20-24,31,6.基金项目
上海市战略前沿专项项目(24DP1501100) （24DP1501100）
