Self-triggered Distributed Model Predictive Control for a Class of Nonlinear MASs under FDI attacks

An aperiodic self-triggered distributed model predictive control algorithm is proposed for discrete nonlinear multi-agent systems (MASs) having sensor-controller channel false data injection (FDI) attacks. An augmented state observer is designed to estimate the real state of the system, and the distributed predictive controller is built on the estimated state such that the MASs stabilize towards the origin. Meanwhile, the self-triggered mechanism is applied to the controller, wherein its triggering condition is derived from the decline of cost function to save computational resources. Then, the recursive feasibility of the optimization problem and the stability of the MASs under the self-triggered distributed model predictive control scheme are analyzed. Eventually, the effectiveness of the proposed method is verified by the simulation results.