Adaptive Neural Finite‐Time Bipartite Containment Control for Stochastic Nonlinear Multi‐Agent Systems With Dynamic Event‐Triggered Mechanism

ABSTRACT In this paper, adaptive output feedback bipartite containment control via a dynamic event‐triggered (DET) mechanism and an improved dynamic surface control (DSC) method is proposed for stochastic non‐strict feedback (NSF) multi‐agent systems (MASs) with unmodeled dynamics and prescribed performance (PP). At present, the definitions of implementing containment control and PP control in stochastic systems are not yet complete. For the first time, two new definitions are given in a probabilistic sense. A high‐gain observer is designed to construct the estimators of unknown states, and a measurable dynamic signal is introduced to deal with unmodeled dynamics. A new prescribed‐time performance function is constructed to reduce the dependence on the initial system values and design parameters, and the constrained bipartite containment error (BCE) is transformed by using a nonlinear mapping. Radial basis function neural networks (RBFNNs) are used to handle unknown continuous functions produced in the recursive process. In previous literature, the assumption that the filtering error is bounded has largely limited practical applications. The improved DSC method not only eliminates the assumption that the filtering error is bounded but also reduces the complexity of controller design. Based on the relative threshold triggering condition and by introducing dynamic variables, the designed DET controller saves more than half of the triggering times compared to the static event‐triggered (SET) mechanism while also avoiding the inclusion of stochastic terms. Finally, through the Lyapunov stability analysis method, it is proved that all signals in the controlled system are semi‐global practical finite‐time stable (SGPFTS) in the sense of moments. In addition, the stability of the proposed control strategy is further verified through two sets of simulation results.