Consensus of a class of multi-agent systems with an active leader under Markovian communication

This paper considers the consensus problem of multi-agent with an active leader under Markovian Communication. Assume the transition probabilities are partly known and the information of the leader can not be gotten by followers directly. To follow such an active leader, the neighbor-based protocol and observer are given for every followers. Then, by applying stochastic stability theory, some sufficient conditions are obtained for multiple agents systems under Markovian communication, under which, the states of each agent can converge to the active leader in the sense of mean square consensus. Finally, numerical simulation results are offered to demonstrate the effectiveness of presented results.