Neuroadaptive fault-tolerant bipartite output formation control for stochastic nonlinear heterogeneous multi-agent systems with multiple leaders and Markovian switching topologies

Abstract

This paper studies fault-tolerant bipartite time-varying output formation (BTVOF) control for stochastic nonlinear heterogeneous multi-agent systems (SNHMASs) with multiple leaders. The interaction topology is signed and switches according to a Markov process, and followers may suffer actuator loss-of-effectiveness faults. We design a distributed leader-state compensator and, for each follower, an output-based observer. Unknown nonlinearities are approximated by adaptive neural networks (NNs) using only local output measurements. Building on these components, we develop a distributed fault-tolerant controller that achieves the desired BTVOF pattern despite stochastic disturbances, nonlinear dynamics, and actuator faults. Rigorous Lyapunov analysis demonstrates that all tracking errors are uniformly ultimately bounded (UUB) in the mean square sense. Simulation results are provided to validate the effectiveness of the proposed control strategy.