面向星地融合网络的星上管控节点动态部署方法

Dynamic Deployment Method of Satellite Management and Control Nodes for Satellite-Terrestrial Integrated Networks

针对低地球轨道卫星网络节点高动态性的问题,提出一种面向星地融合网络的星上管控节点动态部署方法。首先,基于星地融合网络结构复杂多变和星上节点处理能力受限的特点设计主从协同管控的星地融合组网架构;其次,提出星上管控节点部署模型,该模型以星上管控节点部署成本作为优化目标函数,综合考虑业务QoS和网络负载等约束条件,旨在通过优化控制器的数量和位置来改善卫星组网响应时延;最后,采用强化学习方法快速求解不同网络拓扑状态下的星上管控节点数量和部署位置。仿真实验分析了所提算法、加速粒子群算法(ASPO)及最优搜索算法(OptSearch)在星上管控节点动态部署数量和平均组网响应时延上的对比结果,从而验证所提算法在星上管控节点部署中具有一定的优越性。

Aiming at the problem of high dynamics of low earth orbit satellite network nodes, a dynamic deployment method ofsatellite management and control nodes for Satellite-Terrestrial Integrated Networks is proposed.First, based on thecomplex and changeable structure of the Satellite-Terrestrial Integrated Networks and the limited processing capabilitiesof satellite nodes, a Satellite-Terrestrial collaborative network architecture with master-slave collaborative managementand control is designed. Secondly, a satellite management and control node deployment model is proposed. Thismodel takes the satellite management and control node deployment cost as the optimization objective function, andcomprehensively considers constraints such as business QoS and network load. It aims to improve the response delayof satellite networking by optimizing the number and location of satellite control nodes. Finally, the reinforcementlearning method is used to quickly solve the number and deployment locations of satellite control nodes under differentnetwork topology states. The simulation experiment analyzes the comparative results of the algorithm in this article, theAccelerated Particle Swarm Optimization (ASPO) and the Optimal Search (OptSearch) Algorithm in terms of the numberof dynamically deployed control nodes on the satellite and the average network response delay, thereby verifying that thealgorithm in this article has certain advantages in the deployment of on-board control nodes.