基于强化学习的巨型星座离散时间路由策略  

Reinforcement Learning-Based Discrete-Time Routing Strategy for Mega Satellite Constellation

作　　者：储珂 朱立东[1] 成思希 王慧琳 CHU Ke;ZHU Lidong;CHENG Sixi;WANG Huilin(National Key Laboratory of Science and Technology om Communication of UESTC,Chengdu 611731,China)

机构地区：[1]电子科技大学通信抗干扰技术国家级重点实验室,四川成都611731

出　　处：《移动通信》2023年第7期64-70,共7页Mobile Communications

基　　金：国家自然科学基金“卫星隐蔽信号波形设计理论与方法研究”(61871422)。

摘　　要：由于天地一体化网络中移动性和时变环境导致的流量负载、拓扑结构和资源可用性的高时空动态特性,需要考虑当有恶意的网络行为攻击卫星时,如何在未知的网络中探索高效智能的网络优化方案。针对这一问题,以信息年龄、时延抖动率作为多优化目标,提出了一种基于深度强化学习的卫星巨型星座抗干扰路由策略,设置信息年龄、队列增长率以及距离为奖励,实现目标的优化。仿真结果表明,与现有的路由算法相比,该算法不仅可以保证数据的时效性、受到干扰时卫星网络路由的鲁棒性,还可以有效利用星间资源,提升网络的整体性能。The dynamic nature of traffic load,topology,and resource availability in space-ground integrated networks,influenced by mobility and time-varying environments,necessitates the exploration of efficient and intelligent network optimization solutions to combat malicious network behavior targeting satellites in unknown networks.This paper proposes a novel interfcrence-resistant routing strategy for mega satellite constellations,leveraging deep reinforcement learning.The strategy takes into account multiple optimization objectives,including the age of information and delay jtter rate.By employing rewards based on information age,queue growth rate,and distance,the proposed strategy achieves the desired optimization goals.Simulation results demonstrate that the proposed algorithm surpasses existing routing algorithms in terms of ensuring robustness against interference and preserving data timeliness.Additionally,the algorithm ffectively harnesses inter satellite resources,resulting in enhanced overall network performance.

关 键 词：深度强化学习 卫星巨型星座 信息年龄 离散时间动态虚拟拓扑路由策略 

分 类 号：TN915[电子电信—通信与信息系统]