无人集群网络拓扑结构自适应控制技术研究  

Research on Networktopology Adaptive Control Technology of Unmanned Cluster

作　　者：张炜 马殿哲 高天婵 吴日任 刘显 ZHANG Wei;MA Dianzhe;GAO Tianchan;WU Riren;LIU Xian(North Information Control Research Academy Group Co.,Ltd,Nanjing 211153,China;Unit 63966 of PLA,Beijing 100072,China)

机构地区：[1]北方信息控制研究院集团有限公司,南京211153 [2]解放军63966部队,北京100072

出　　处：《火力与指挥控制》2022年第12期152-156,共5页Fire Control & Command Control

基　　金：军委科技委重大专项基金资助项目(xxxx20002)。

摘　　要：为满足复杂对抗环境下无人集群对可靠协同信息处理的基础网络需求,降低因任务多样、通信能力有限、环境未知、战场条件恶劣等导致系统网络拓扑结构断裂的概率,提高无人集群系统的网络容错性和可靠性,设计了网络拓扑结构自适应控制方法,即适应于复杂环境和任务状态多变的动态拓扑结构调整技术。该技术通过QMIX算法实现集群的网络重构,利用多智能体自适应调整策略,在保持集群连通拓扑结构的同时,渐进实现拓扑结构总体发现时间最小和组网能耗最优。In order to meet the basic network requirements of unmanned cluster for reliable cooperative information processing in complex confrontation environment.Reduce the probability of network topology fracture of unmanned cluster system caused by diverse tasks,limited communication capacity,unknown environment and bad battlefield conditions,and improve the network fault tolerance and reliability of unmanned cluster system.The adaptive control method of network topology is designed,which is a dynamic topology network adjustment technology suitable for complex environment and changeable task state.QMIX algorithm is used to realize the network reconfiguration of the cluster,and multi-agent adaptive adjustment strategy is used to keep the connected topology of the cluster while gradually realizing the minimum overall discovery time and the optimal network energy consumption.

关 键 词：网络拓扑结构 自适应控制 多智能体 无人集群 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]