Fully distributed and attack-immune protocols for linear multiagent systems by linear time-varying feedback

Abstract

This paper addresses the leader-following consensus problem for general linear multiagent systems under general directed topology, with a focus on the scenarios where only relative output information is available. To solve such problem, most existing observer-based protocols rely on relative observer information among neighboring agents, which is obtained through the communication network and can be susceptible to network attacks. To overcome this limitation, this paper proposes a novel class of distributed observer-based linear time-varying protocols, in which the feedback gain is designed with two time-varying terms, one of which tends towards infinity as time increases, while the other tends towards zero. Compared to existing protocols, the proposed protocols offer significant advantages. Notably, they do not require information exchange through the communication network, making them immune to network attacks. Furthermore, the protocols operate in a fully distributed manner, which makes them resilient to changes in the topology graph. Simulation results demonstrate the effectiveness of the proposed approach.