Fault detection and time-varying formation control for nonlinear multi-agent systems with Markov switching topology

This paper introduces a co-design approach for fault detection and time-varying formation control for a series of nonlinear leader-following multi-agent systems under stochastic switching directed topology. The random environmental fluctuations are described by a Markov chain. Utilizing relative output information, fault detection observers are developed not only to estimate system states but also to construct a robust residual signal, capable of detecting multiple faulty agents simultaneously in the presence of external perturbations. The paper provides sufficient conditions to guarantee the uniform ultimate boundedness in probability of the formation error and the H-infinity performance for fault detection. Precise identification of potential faulty agents is achieved by evaluating the residual signal and logical reasoning. The effectiveness of formation control and the efficiency of the proposed joint design approach are demonstrated by a set of tailless fighter airplanes.