Reduced-Order Output-Feedback Bipartite Consensus of Linear Multi-Agent Systems

Abstract

This paper investigates state bipartite consensus of homogeneous linear multi-agent systems based on output feedback. A reduced-order observer-type dynamic output-feedback protocol is proposed. Solvability conditions for the bipartite consensus problem under the proposed protocol are derived. Analytic expressions of the consensus state are obtained. Compared with existing results, the proposed protocol has two innovations: first, the protocol relies on relative information for in-group agents and, second, the protocol state is ensured to converge to zero even if the agent states are divergent. Numerical examples are finally presented to demonstrate the effectiveness of the designed control protocol.