无人机-无人车跨域编队跟踪的分布式博弈策略  

Distributed Game Strategy for Formation Tracking of UAV-UGV Cross-domain Swarm System

作　　者：石哲鑫 冯永进 于江龙 董希旺[1,2,3] 李清东 任章[1] SHI Zhexin;FENG Yongjin;YU Jiangong;DONG Xiwang;LI Qingdong;REN Zhang(School of Automation Science and Electrical Engineering,Beihang University,Beijing 100191,China;Institute of Artificial Intelligence,Beihang University,Beijing 100191,China;Institute of Unmanned System,Beihang University,Beijing 100191,China;Shenyuan Honors College,Beihang University,Beijing 100191,China)

机构地区：[1]北京航空航天大学自动化科学与电气工程学院,北京100191 [2]北京航空航天大学人工智能研究院,北京100191 [3]北京航空航天大学无人系统研究院,北京100191 [4]北京航空航天大学沈元荣誉学院,北京100191

出　　处：《无人系统技术》2024年第2期28-38,共11页Unmanned Systems Technology

基　　金：国家自然科学基金(62103016,U2241217,61973013,62103023);北京科协市青年人才托举计划(BYESS2022186);北京市自然科学基金(4232046);中国科协青年人才托举计划(2021QNRC001)。

摘　　要：针对无人机-无人车跨域系统的编队跟踪与扰动抑制问题,开展了领-从架构下考虑存在扰动的跨域编队跟踪的分布式博弈策略研究。首先,建立了含扰动的无人机-无人车编队跟踪模型,每架无人机只使用局部信息来最小化自己的性能指标,而整个跨域系统试图寻求基于通信拓扑的最优决策。其次,提出了一种最差扰动下的博弈策略,并严格证明了系统中存在纳什均衡。然后,引入了一个分布式终端状态观测器,得到了所提出博弈策略的局部实现。最后,搭建了一套无人机-无人车跨域系统验证平台,并通过一个跨域编队跟踪应用案例验证了所提出分布式博弈策略不仅可以有效地抑制扰动,还能形成期望编队队形,并且性能指标在数值仿真和实物试验下的差异不超过0.6%,证明所提出分布式博弈策略有良好的实用性。A distributed game strategy is studied to address the formation tracking and the disturbance rejection issues of unmanned aerial vehicle(UAV)and unmanned ground vehicle(UGV)cross-domain systems based on the leader-follower structure considering external disturbances.Firstly,a UAV-UGV formation tracking model with disturbances is established.Each UAV uses only local information to minimize its own performance index,while the entire cross-domain system attempts to seek an optimal decision based on the communication topology.Secondly,a game strategy under the worst-case disturbance is proposed.The existence of Nash equilibrium in the cross-domain system is rigorously proved.Then,a distributed terminal state observer is introduced to obtain a local implementation of the proposed game strategy.Finally,a UAVUGV cross-domain system verification platform is established.The effectiveness of the proposed distributed game strategy is verified by a cross-domain formation tracking application case,which not only effectively rejects disturbances,but also forms the expected formation configuration.The difference in the performance index between the numerical simulation and the physical experiment does not exceed 0.6%,which also reflects the practicability of the proposed distributed game strategy.

关 键 词：跨域系统 编队跟踪 扰动抑制 博弈策略 纳什均衡 分布式观测器 无人机-无人车 

分 类 号：V279[航空宇航科学与技术—飞行器设计] TP29[自动化与计算机技术—检测技术与自动化装置]