Consensus tracking for stochastic multi-agent systems with actuator faults and DoS attacks: mode-dependent average dwell time strategy

This article investigates the consensus tracking problem for stochastic multi-agent systems with noise disturbances, actuator faults, and switched topologies induced by Denial-of-Service attacks. A key challenge lies in the coupling of noise and actuator faults, which may cause state divergence during connectivity-broken attacks. To address this issue, a novel fault-tolerant control protocol is proposed based on a mode-dependent average dwell time strategy. Compared with conventional approaches, the proposed protocol not only achieves a tighter error upper bound and faster convergence, but also removes the restrictive upper bound constraint imposed on the average dwell time during connectivity-broken attacks. Furthermore, topology-specific control gains are introduced to enable more precise fault compensation. The performance of the proposed protocol is validated via a numerical simulation and an application to secondary voltage control in microgrids.