基于态势演化博弈的无人机集群动态攻防  

Dynamic offense and defense of UAV swarm based on situation evolution game

作　　者：盛磊 时满红[1] 亓迎川[1] 李浩[1] 庞明军 SHENG Lei;SHI Manhong;QI Yingchuan;LI Hao;PANG Mingjun(Air Force Early Warning Academy,Wuhan 430000,China;Unit 95894 of the PLA,Beijing 100000,China)

机构地区：[1]空军预警学院,湖北武汉430000 [2]中国人民解放军95894部队,北京100000

出　　处：《系统工程与电子技术》2023年第8期2332-2342,共11页Systems Engineering and Electronics

基　　金：国家自然科学基金(61502522)资助课题。

摘　　要：针对无人机集群动态攻防问题,提出了态势演化博弈模型。首先,从基地与守方无人机集群相互作用以及攻防双方动态斗争角度出发,构建了攻方无人机与守方无人机的态势演化博弈模型。其次,设计了同态势演化博弈模型相对应的态势评估函数,使每阶段个体的策略选择都为该阶段的最优解。然后,根据策略选择,结合社会力模型,驱动无人机向指定目标运动,完成攻防双方无人机集群的动态对抗。实验结果表明,无人机集群的动态运动符合空战情况,验证了基地在具备功能时的优越性,以及设计的态势演化博弈模型能够实现无人机集群对抗决策的自适应选择。To solve the problem of dynamic offense and defense in unmanned aerial vehicle(UAV)swarms,a situation evolution game model is proposed.Firstly,building upon the interactions between both the base and the defender’s UAV swarms,as well as the descriptions of the dynamic struggle between offender and defender,a situation evolution game model is established by taking into account situation evolution among offending and defending UAVs.Secondly,a corresponding situation evaluation function is designed to optimize individual strategy selection at each stage of combat.Then,based on this strategy selection and through the use of the social force model,the UAVs are driven to move towards designated targets,resulting in a dynamic confrontation between the offending and defending UAV swarms.Experimental results demonstrate that the dynamic movement of the UAV swarms conforms to those observed in actual air combat scenarios,verifies the superior position of a functioning base,and confirms the accuracy of the UAV swarms’adaptive decision-making through the use of the situation evolution game model.

关 键 词：无人机 演化博弈 态势评估 智能体决策 集群对抗策略 

分 类 号：V279[航空宇航科学与技术—飞行器设计]