A distributed FDI cyber-attack detection in discrete-time nonlinear multi-agent systems using neural networks

This paper proposes a distributed, false data injection (FDI) cyber-attack detection method in communication channels for a class of discrete-time, nonlinear, heterogeneous, multi-agent systems controlled by our formation-based controller. A distributed neural network (NN)-based observer is proposed that generates the residual signal which is used in detection of FDI attacks on agents’ sensors, actuators, and neighboring communication channels in a multi-agent formation control setting. A radial basis function neural network (RBFNN) is used to approximate the unknown nonlinearity in the dynamics. A Lyapunov stability theory is used to prove that the attack detection residual and the multi-agent formation error are uniformly ultimately bounded (UUB), and to explicitly derive the NN weights tuning law and the attack detectability threshold. The proposed method’s attack detectability properties are analyzed, and simulation results are provided to demonstrate performance of the detection methodology.