Output Consensus Tracking of Heterogeneous Open Multi-Agent Systems Under Actuator Attacks

This paper investigates the output consensus tracking problem of a class of heterogeneous open multi-agent systems (oMASs) under actuator attacks. The main challenges in solving this problem stem from the presence of actuator attacks and the dimensionality mismatch of the error signals caused by the dynamic entry and exit of followers. To address such challenges, a novel analytical method is developed to characterize the error evolution at switching instants via switching-instant error-coupling inequalities, which describe how error signals evolve in response to changes in the number of followers. Based on this method, a full-order attack-free state observer is designed to estimate the states of each follower without estimating the exact value of the attack. In addition, a reduced-order attack-free observer is developed to reduce the estimation burden and computational complexity, while still ensuring the effective estimation of the states. Building on the observer architectures, both full-order and reduced-order observer-based distributed resilient control protocols are proposed to achieve output consensus tracking while ensuring the tracking error remains uniformly ultimately bounded (UUB) for heterogeneous oMAS under actuator attacks, by employing the output regulation approach. The effectiveness of both types of control protocols is demonstrated through numerical simulation.