Distributed mean-square consensus of heterogeneous multi-agent systems with packet losses

Abstract

This paper studies the mean-square consensus problem of discrete-time multi-agent systems, where the state information is received with different uncertainty of dropout which leads to the heterogeneous observation. Different from previous consensus studies without considering measurement packet dropout, the main difficulties encountered in this paper are the controller design for each agent with different dynamics and the state estimator design of stochastic systems with multiplicative noises. First of all, the optimal estimator is designed based on the measurement to estimate the agent’s state information, and a distributed dynamic output feedback control protocol is presented based on the obtained estimator. Secondly, applying the distributed feedforward control method and constructing the stochastic Lyapunov-Krasovskii functional, the sufficient solvability conditions for heterogeneous multi-agent systems achieving mean-square consensus is derived for the first time. Moreover, the stability of the error covariance matrices related with the optimal estimator is studied. Finally, a numerical example is provided to illustrate the effectiveness of the proposed algorithm.