Distributed multivariate-observer-based robust consensus control of nonlinear multiagent systems against time-varying attacks on actuators and sensors

This article investigates the robust consensus problem for nonlinear multiagent systems against time-varying false data injection attacks on actuators and sensors. First, the root-mean-square (RMS) theory is used to extend the assumption of the slow-varying or constant attack signals to the case of time-varying attack signals. Second, a novel distributed multivariate observer (DMO) is designed to estimate the followers' system states and the time-varying attack signals on actuators and sensors. With the help of the outputs of DMO, a distributed robust consensus control arithmetic is proposed, which can compensate for actuator attacks and isolate sensor attacks so that exponential consensus and robust consensus are achieved. In particular, the robust performance of estimation errors and consensus errors is ensured by establishing the RMS gain index via linear matrix inequality, in which the zero initial conditions of estimation errors and consensus errors are not required. Finally, two simulation examples, including a network of four aircraft longitudinal dynamic systems, are given to verify the effectiveness of the proposed arithmetic.