Attack reconstruction-based resilient consensus tracking for multi-agent systems with distributed adaptive control protocol

This article considers the distributed secure consensus tracking control problem for multi-agent systems (MASs) with relative output information. To tackle the potential false-data-injection (FDI) attacks on both the communication between physical components and the control center at the agent level, as well as the message transmission between connected agents, local sliding mode observers (SMO) and distributed observers are proposed for each agent. By introducing reconstructed signals for attacks and smooth terms to compensate the unknown input in the leader agent's dynamics, a distributed adaptive consensus protocol with relative output information is proposed. The closed-loop stability is analyzed by introducing a novel Lyapunov function. It is shown that all closed-loop signals are globally uniformly bounded and asymptotically consensus tracking can be achieved. Simulation results are provided to validate the theoretical findings.