Collective tracking control for multi-agent system on balanced graphs

In this paper, we consider a collective tracking problem of a multi-agent system which has a varying-velocity leader and evolves on balanced information topologies. First, we propose a coordination-variable-based control strategy for every agent to track the leader whose state may not be measured. Moreover, from a collective behavior perspective, a new decomposition method is utilized to decompose the resulting closed-loop system into two subsystems, and a novel measure also is introduced to characterize collective tracking performance of the system. Finally, we prove that the proposed control strategy can not only estimate the tracking errors but also increase the introduced measure to improve collective tracking performance even under the case of balanced information topologies.