Semi-Global Robust And Output Consensus Of Multi-agent Systems With Input Saturation

Consensus problem of multi-agent systems,as a fundamental problem in distribut-ed cooperative control,has drawn much attention from the control community.The the-ory of consensus has been widely applied in many fields,such as cooperative formation control of multiple robots,cooperative motion control of multiple vehicles,distribut-ed attitude control of multiple spacecrafts and distributed information fusion of sensor networks.Consensus of multi-agent systems means that all agents reach a common value under a distributed protocol based on local information exchange.According to the number of leader agent,the consensus problem can be classified as the leader-less consensus problem,the leader-following consensus problem and the containment control problem.In the heterogeneous multi-agent system that the leader agent and fol-lower agents are with different dynamics,follower agents need to track the reference and/or reject the disturbance,where the reference and the disturbance are generated by the leader agent.The consensus problem of the heterogeneous multi-agent system is output consensus problem,which can be classified as the leader-following output consensus problem and the output containment problem,according to the number of leader agent.In view of the ubiquity of input saturation in practical system,the consensus prob-lems with input saturation are of certain significance both in theoretical research and in engineering applications.In this dissertation,we consider two kinds of actuator nonlin-earities,the standard saturation function and the imperfect actuator.Based on the low gain feedback protocols,we investigate the semi-global consensus problem,the robust semi-global consensus problem and the semi-global output consensus problem of the multi-agent systems with input saturation.The main contribution of this dissertation is stated as follows1.Robust semi-global consensus of multi-agent systems with imperfect actuators.The imperfect actuators are characterized by nonlinearities such as saturation and dead zone and their input output relationships are not precisely known.The dynamics of agents are also affected by the input additive disturbances.Two consensus protocols,of the low-and-high gain feedback type and the low gain based variable structure con-trol type,are proposed to solve the semi-global leader-following consensus problem of the multi-agent system with imperfect actuators over a strongly connected and de-tailed balanced communication topology.The robust semi-global leaderless consensus(robust semi-global containment control)of the multi-agent system with imperfect ac-tuators can be achieved by the low-and-high gain feedback consensus protocol over an undirected networks among the agents(follower agents).2.Semi-global consensus of discrete-time multi-agent system with input satura-tion.The actuator nonlinearity can be represented by a standard saturation function.A parametric ARE is modified,based on which a low gain consensus protocol is pro-posed to solve the semi-global leader-following consensus problem,the semi-global leaderless consensus problem and the semi-global containment control problem of the discrete-time multi-agent system with input saturation over directed networks.3.Semi-global output consensus of heterogeneous multi-agent systems with input saturation.To make the output of the follower agents track that of the leader agent,the information of the leader agent is required by each follower agent for controller design.But not all of the follower agents have access to the information of the leader agent.To overcome this difficulty,we construct a distributed leader state observer to estimate the leader state for each follower agent.Based on the estimates of the leader state,low gain consensus protocols are proposed to solve the semi-global leader-following output consensus problem and the semi-global output containment control problem of the heterogeneous multi-agent system with input saturation.4.Semi-global output consensus of heterogeneous discrete-time multi-agent sys-tems with input saturation.The mechanism of high gain distributed leader state ob-servers for continuous-time does not apply in the discrete-time setting.Thus,we con-struct a discrete-time distributed leader state observer,which does not rely on high gain action,to estimate the leader state for each follower agent.Based on the estimates of the leader state,we design low gain consensus protocols to solve the semi-global leader-following output consensus problem and the semi-global output containment control problem of the heterogeneous discrete-time multi-agent system with input saturation.