考虑未知输入的异构集群系统群体智能合围跟踪控制  

Formation-containment tracking intelligent control for heterogeneous swarm systems under unknown input

作　　者：王林波 王蒙一[1] 周思全 江涌 化永朝 董希旺 WANG LinBo;WANG MengYi;ZHOU SiQuan;JIANG Yong;HUA YongZhao;DONG XiWang(Beijing Instituteof Electronic System Engineering,Beijing 100854,China;Changfeng Electromechanical Technology Design Academy,Beijing 100854,China;Instituteof Artificial Intelligence,Beihang University,Beijing 100191,China;School of Automation Science and Electronic Engineering,Beihang University,Beijing 100083,China)

机构地区：[1]北京电子工程总体研究所,北京100854 [2]中国长峰机电技术研究设计院,北京100854 [3]北京航空航天大学人工智能研究院,北京100191 [4]北京航空航天大学自动化科学与电气工程学院,北京100083

出　　处：《中国科学：技术科学》2023年第2期291-306,共16页Scientia Sinica(Technologica)

基　　金：科技创新2030-“新一代人工智能”重大项目(编号:2018AAA0102305)资助。

摘　　要：集群系统合围跟踪控制是群体智能涌现在运动控制层面的实现途径之一,在智能化战争时代具有广阔的应用前景.本文面向未来空地群体智能作战中护卫任务的应用需求,聚焦军品押送、重要人物护送等典型任务场景,研究了异构集群系统考虑领导者具有未知输入的分布式编队-合围跟踪控制问题,以实现复杂环境下空地协同智能作战.首先引入代数图论知识,建立具有未知输入的异构集群系统模型,提出了一种编队构型生成系统.之后结合自适应控制律,利用邻居局部信息交互设计了基于边的分布式编队-合围跟踪控制器,并证明了闭环系统的稳定性.此外,对异构集群系统涌现出的更复杂的群体智能特征进行了分析与总结.最后,通过数值仿真验证了本文所提出的方法能够实现预期的编队-合围跟踪控制,并建立了实物等效验证系统,对控制方法的有效性进行了验证,为实际情况下大规模空地协同智能作战提供了有力的理论支撑.The containment tracking control of swarm systems is an important approach for the emergence of swarm intelligence at the control level, which has potential applications in the intelligent war era. Aiming at the requirements of military tasks in the future battlefield and focusing on typical mission scenarios, such as military goods and important people escort, this thesis investigates the distributed formation-containment tracking control problem of heterogeneous swarm systems under an unknown input to realize intelligence emergence in air-ground cooperative operation. First, the algebraic graph theory is introduced to establish a hierarchical heterogeneous swarm system model with unknown input. Next, in combination with the adaptive control law and by using the local information interaction between neighbors, the edge-based distributed formation-containment tracking controllers are designed, and the stability of the closed-loop swarm systems is proven;otherwise, the more complex swarm intelligence characteristics in this thesis are analyzed and summarized. Finally, a simulation example is provided to demonstrate that the proposed approach can achieve the expected formation-containment tracking control. A physical equivalent verification system is also established to verify the effectiveness of the approach, which provides strong theoretical support for the large-scale cooperative escort of air-ground.

关 键 词：异构集群系统 合围跟踪控制 未知输入 群体智能 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置] E91[自动化与计算机技术—控制科学与工程]