Research On Cooperative Control Of Multi-agent Systems Under Switched Topologies

As the rapid development of intelligent control systems,more and more new issues in this field are proposed.One of them is the cooperative control of multi-agent systems.Cooperative control of multi-agent systems means a group of agents achieve a global goal which only relies on each agent making its own decision referring to the information of its own and its neighbors'.There are many kinds of cooperative control of multi-agent systems referring to the categories of cooperative control or the network of systems.The consensus is one basic problem of cooperative control of multi-agent systems.More and more problems have been resolved for the cooperative control of multiagent systems under fixed topologies,while the study under switched topologies is getting on slowly.Switching topologies are more complex than the fixed topologies.Studying switching topologies is challenging.This dissertation studies the cooperative control of multi-agent systems under switched topologies.Its main contribution is:(a)proposing a topology-dependent average dwell time method to study the multiagent systems with improper parameters or under topologies with no spanning trees.In this dissertation,it is proved that the second-order or third-order or higher-order systems can reach a consensus as long as the conditions satisfy the theorems in this dissertation even if some topologies are with improper parameters or with non-spanning tree topologies.(b)proposing a delta method to construct a unified framework for continuous-time,discrete-time and sampled multi-agent systems under switched topologies with faults.The cases that some systems have no spanning tree or improper parameters are also discussed.It is proved that multi-agent systems with faults can reach a consensus so long as the conditions satisfy the theorems in this dissertation.(c)analyzing the multi-agent systems with asynchronous switching.There is little discussion on the asynchronous switching for multi-agent systems in the literature.An example illustrates that many multi-agent systems are far away from the commercial goal when asynchronous switching occurs.The problems of the multi-agent systems with asynchronous switching are analyzed and some sufficient conditions are proposed.(d)researching consensus,tracking consensus and containment control and their relationship.In this dissertation,a unified framework and an error index based on local information are built.The proposed results are fit for both balanced and unbalanced networks.