Design Of Consensus Control Protocols For Multi-agent Systems With Time-delay

The multi-agent system has been regarded as a leading edge discipline in the control field gradually, due to their broad applications in science and engineering. Distributed control and local control are vital branches of the multi-agent system. The idea of distributed control is decomposing a complex task into several simple tasks which are assigned to agents. The multi-agent system can reach the consensus point by using the distributed controllers. The purpose of the local control is to drive the multi-agent system to reach consensus by adding the local controller into single or several agents. In this thesis, the distributed control strategy for the multi-agent systems with time delay is proposed. Moreover, the local control scheme that improves the convergence speed of multi-agent system is presented.In the research of the distributed cooperative control, the consensus problem of the second-order multi-agent systems can be transfomed into the problem of obtaining the stability region of the PD control parameters. The stability regions of the PD control parameters are presented on the basis of the Routh stability criterion for the second-order multi-agent system free of time delay. The stability regions of the PD control parameters are presented by the D-decomposition method for second-order multi-agent system with time-delay. The second-order multi-agent system can reach consensus when the parameters of PD controller are selected from the resultant stability regions. The cooperative control strategy for the third-order multi-agent system with time delay is more complicated than the one for the second-order multi-agent system. The method of obtaining the stability regions of the PID controller is presented for the third-order multi-agent system with time delay. Firstly, the stability range of the proportional gain is determined, then, the stable region of the integral and differential parameters in the two-dimension plane is given for a fixed proportional gain. The stability region of the distributed PID control protocol has the linear programming property.In the research of the fast consensus under the local control scheme, the consensus of the mutli-agent system is need to be guaranteed when the local controller is addded into any single agent. Firstly, select the optimal agent that can drive the multi-agent system to reach consensus point fastest based on the properties of the node. Then the range of the controller parameter is obtained by transfer function of every agent. After that, analyze the consensus and the convergence speed of the multi-agent system. Simulation examples show the effectiveness of the algorithm.