Research On Output Regulation And Consensus Of Multi-Agent Systems

With the rapid development of the sensor technology and embedded chips, distributed multi-agent control systems have gradually become a hotspot of automation engineering and control theory. Closed-loop feedback control system via wireless sensors and wireless communication is called a multi-agent system. Compared with the traditional point to point control system, multi-agent systems are famous for high reliability, low cost, easy to configure and maintain and other advantages, which have been widely used. However, since the agent has limited energy, limited sensor and limited microchip and many other restrictions, it is always quite complicated to the analysis and design of multi-agent control system.Firstly, for linear cooperative output regulation of multi-agent systems, the problem about containment control is studied in this thesis. Raw output regulation problem originated in a single leader follower multi-agent system, will be derived from the single leader and several followers to more leaders and these followers. Containment control law can ensure that all followers can track the trajectory leader and the output error gradually approach zero when time elapses. Lyapunov stability theory is used to design a control algorithm to drive all the followers to enter the convex hull formed by the leader. The controller design procedure is given to determine the feedback gain matrix and the appropriate gain matrix is selected to make ones achieve the consensus.Secondly, for the consensus problem of the second-order multi-agent systems. distributed event-triggered strategy is used in the thesis. This strategy is based on the norm ratio between the predefined measurement error and the state equation to update the control input. According to the equations of the related agent, its event trigger needs the information of its own and their neighbors. The procedure of the distributed event-triggered strategy is given by Lyapunov stability theory and Linear Matrix Inequality (LMI) approach for a second-order multi-agent systems to achieve the purposes of convergence.Finally, for a set of multi-agent systems, self-triggered consensus strategy is designed. The next time when the agents update its control input is calculated by each agent at a predefined time. This strategy extends the current work of the event-triggered strategy, as for the self-triggered strategy, it is not necessary for the agent to keep track of the state so as to continuously trigger the control actuator in the update interval. Lyapunov stability theory and LMI method is used for a particular first order agent to realize the goal of convergence.