Task-oriented Fault-tolerant Topology And Flocking Control For Multi-agent Systems

With the development and cross fusion of the control science and sociology, artificial in-telligence, computer science, biology, distributed cooperative control of multi-agent systemsand application is more popular recently and has gradually become a hot topic in field ofcontrol science. Each agent achieves the overall planning and solves the local conflicts, andthen achieves the expected global task by using local information. The multi-agent system hasimportant theoretical research value and broad application prospects in the industrial and mil-itary as a comprehensive subject. Flocking control of multi-agent systems is a representativeproblem in the field of multi-agent coordination, and is also the basis of many other decentral-ized control and estimation problems. On the one hand, the topology of multi-agent systemschanges dynamically, and the node has limited energy. The constantly changing topologywill influence the performance of the system since some nodes have crashed even though thesystem is operating normally. On the other hand, the multi-agent systems encounter compli-cated, harsh and difcult environment. The topology may sufer network partition becauseof some factors such as unknown environment,disgusting condition, diversity of tasks, andlimited communication capacity and perception, the the cooperative control task would faileventually. It is necessary to enhance the fault tolerance of the topology for improving the co-operative capability of multi-agent systems in complex environment. On the basis of reviewand summary of the corresponding research, the thesis studies the fault-tolerant topology ofmulti-agent system the flocking problem by using the graph theory and matrix theory. Themajor contributions of the thesis are as follows:(1) Considering the traditional planar topology which is lack of flexibility in the disgust-ing environment and is unable to describe the mass group, a hierarchical model and afault-tolerant biconnected topology is constructed based on the hierarchical network.Finally, the simulation results show that the proposed algorithm achieves the flexible,reliable and self-healing topology which guarantees the transmission of information timely in the background of networked air defense combat.(2) With the aim of constructing self-healing topology, a K-connected topology controlalgorithm that can cope with faults such as node failures and link disruptions is pro-posed. The system stays connected in the dynamic interaction topology, even in theface of K1nodes departure. Our approach combines power transmission and mo-tion control for constructing a K-connected topology with approximately minimumpower to prolong their working life which only use local information. Finally, severalnumerical simulations are provided to illustrate the efciency of the algorithm on thebasis of the topology of the command system and executive system of the networkedfire control system for field air defense is vital in air defense combat environment.(3) The problems of flocking with both connectivity maintenance and obstacle avoidancefor the system of dynamic agents are addressed. In the case where the initial topol-ogy is connected, a decentralized flocking control protocol is proposed to enable thegroup to asymptotically achieve the desired stable flocking motion using artificial po-tential functions combined with stream functions, which could not only maintain thetopology connectivity of the dynamic multi-agent systems for all time but also makeall the agents avoid obstacles smoothly without trapping into local minima. Finally,the simulations and physical experiment of multi-robot system are presented to verifythe efectiveness of the algorithm and application value in the networked fire controlsystem for field air defense on the basis of the executive system of the networked firecontrol system for field air defense has the requirement of flocking task with obstacleavoidance.(4) The fault-tolerant topology control problem of agent groups in the context of multi-agent flocking tasks with second order non-linear dynamics is investigated. By intro-ducing the concept of biconnectivity as the basic requirement of fault-tolerant topolo-gy, a simple fault detector, a motion control strategy of biconnectivity and a flockingalgorithm with bounded control input are proposed. It is proved that the proposedcontrol algorithm can not only achieve the resultant biconnected topology which is able to tolerate temporary node failures, but also guarantee the stable flocking motion.Thus, the combination of the fault-tolerant topology and flocking control is achieved.Finally, several numerical simulations are provided to illustrate the efciency of thealgorithm by considering the requirement of the task-oriented fault-tolerant topologyfor the networked fire control system for field air defense in complicated and difcultenvironment.