Resilient Consensus Control for Heterogeneous Multi-Agent Systems Subject to Two-Channel Stochastic Deception Attacks

This paper explores the consensus control for heterogeneous multi-agent systems (MASs) subject to two-channel cyber-attacks over a directed network. A consensus control is a cooperative control scheme in which a group of multiple systems works collaboratively to reach similar final states in a finite time. Deception attacks targeting the input signal and the connections between agents have been considered by incorporating stochastic addition of the attack signals. The considered deception attacks are more complicated to deal with than conventional false data injection, owing to consideration of both blockage of actual signal and injection of false data. A state feedback control scheme using a dynamic estimator is proposed to achieve practical consensus among follower agents and the leader. Relevant conditions have been provided for attaining the stability and for designing the resilient controller for heterogeneous MASs to keep the consensus error uniformly ultimately bounded (UUB). Incorporating heterogeneous MASs, two-channel cyber-attacks, and stochastic nature of deception attacks has rendered the presented methodology more sophisticated, practical, and challenging compared to existing methods. Additionally, the proposed results have been expanded for the output feedback consensus through the incorporation of a Luenberger observer. In this context, a decoupling strategy has been utilized to ensure consensus among heterogeneous agents through resilient output feedback, considering stochastic attacks targeting two channels. The results of the proposed resilient consensus control for heterogeneous agents are verified through numerical simulations of circuits.