Study On Consensus Of Multi-agent Systems With Markovian Switching Topologies

In recent years,due to the rapid development of communication network,the application of multi-agent system has become more and more extensive.However,there are still some shortcomings.As we all know,in the process of network transmission,the phenomena of delay,packet loss and disorder are unavoidable,which will lead to system performance degradation.Most literatures adopt periodic sampling method to update the controller,but this method will lead to unnecessary data transmission.A new trigger mechanism,event-triggered,emerges as the times require.Compared with the traditional periodic triggering mechanism,event-triggered can effectively reduce the number of updates of the controller and system power consumption.In this thesis,the multi-agent systems is used as the carrier,and the consistency and filtering problems of the system based on event-triggered mechanism are analyzed emphatically.The main work of thesis is as follows:1.Leader-following consensus of nonlinear multi-agent systems with Markovian switching topology is studied.Considering the influence of network-induced delay on the whole system,a unified model of the system is established by time-delay partitioning method.By constructing Lyapunov functional and combining LMI technique,sufficient conditions for asymptotic stability of closed-loop systems are obtained.The conservativeness of the system is reduced by introducing the free weight matrix method,but the computational complexity is increased.2.The controlled object is the multi-agent systems with Markovian switching topology.The H_?filtering problem based on this model is discussed.At the same time,we introduce a distributed event-triggered mechanism to reduce the traffic of network communication.By constructing Lyapunov function and adopting time-delay partitioning method,sufficient conditions for asymptotic stability of filtering error system are obtained,and H_?filters are further designed.An interactive convex combination method based on Jensen inequality is used to deal with cross terms,which greatly reduces the computational complexity of the filtering error system.