Secure Consensus of Multiagent Systems via Impulsive Control Subject to Deception Attacks

Abstract

This brief studies the secure consensus problem of multiagent systems (MASs) via impulsive control under deception attacks. In most existing work, deception attackers are assumed to attack the agents rather than the communication channels, and the neighbors of the attacked agent receive the same tampered information from the attacked agent. The deception attacker is assumed to inject false data into the communication channels, and each neighbor of each agent may receive different tempered information from the same agent. Attacking the communication channels leads to an inconsistent dimension of the matrices in the controller, which brings difficulties in analyzing the convergence of MASs. To overcome the challenge, the extended diagonal matrix is proposed to analyze the effect of the deception attacks on the convergence of MASs, and sufficient conditions for the secure consensus are derived. The consensus error bound converges to a bounded area dependent on the total attack energy. Finally, a simulation example is given to verify the effectiveness of the results.