Study On Formation Control Of Multi-agent Systems Based On Consensus Protocol

In recent years, the coordination control of multi-agent systems has received extensive attentionof the scientific community, and is used in satellite flying formation, unmanned aerial vehiclestracking. The key problem of the multi-agent systems is how to put forward a kind of distributedcontrol strategy based on local information, in order to make all the agent’s state informationachieve consistent, namely consensus problem.In this paper, the second-order consensus problems for systems modeled by second-orderdynamics are mainly discussed. To make the state information of multi-agent systems converge to acommon value, the suitable consensus algorithm is proposed. Besides, the consensus algorithm isvaried and applied to the formation control of multi-agent systems. The main research work andachievement are as follows:First, for second-order dynamic model of multi-vehicle system, we investigate a second-orderconsensus protocol for the system in the case of collecting information without time-delay. Byintroducing position tracking value, we propose a simple second-order consensus protocol, makingthe state of each vehicle converge to the given tracking values. The necessary and sufficientcondition under which second-order consensus can be reached is derived by using the systemstability theory. The second consensus is then applied to the multiple vehicle systems with fourvehicles as an illustrative example.Second, we propose a simple consensus algorithm for systems modeled by second-orderdynamics, and derive necessary and sufficient conditio ns under which consensus can be reachedbased on algebraic graph theory and matrix analysis. We apply the simple second-order consensusalgorithm to tackle formation control problems. The second-order consensus protocol and twomulti-vehicle systems formation control example are shown in simulation to illustrate our strategies.Third, both the standard position deviation and the relative position deviation are considered,the second-order consensus protocol based on the position difference is proposed. Still, using matrixanalysis and the knowledge of graph theory, the stability of the proposed consensus protocol isproved. It is worth noting that comparing with the second research point, the convergence speed ofthis algorithm is faster. The second-order consensus protocol and a multi-vehicle system formationcontrol example are shown in simulation to illustrate our strategies.