Consensus Of Multi Agent Systems Based On Leader Following Control With Measurement Noise

Consensus of MAS is a very important problem in complex network research. In this paper, a class of multi agent systems based on leader control model is discussed. We assume the coupling topology of the attraction action among the agents is different from the coupling topology of the repulsion action. The information that each agent received is corrupted by measurement noises. By using the tools of stochastic analysis and algebraic graph theory, a sufficient condition is obtained for the protocol to ensure consensus under fixed topologies. It is proved that the agents will converge and form a cohesive cluster following the leader and finally enter and remain in a bounded hyper-sphere. Furthermore, by using a common Lyapunov function, the result is extended to the switching topology case.