Dynamic Event‐Triggered Adaptive Neural Finite‐Time Containment Control for Switched Stochastic Multi‐Agent Systems With Prescribed Performance

ABSTRACT In this paper, adaptive neural output feedback finite‐time containment control is presented based on improved dynamic surface control (DSC)for nonstrict‐feedback (NSF) switched stochastic multi‐agent systems (SMASs) with event‐triggered (ET), prescribed performance (PP), and unmodeled dynamics (UD). Two novel definitions are put forward to achieve containment control and fulfill the PP constraints in switched SMASs. High‐gain observers are utilized to construct the estimators of unknown states, and measurable dynamic signals are introduced to deal with UD. By using the hyperbolic tangent function and nonlinear mapping, the containment error can reach a prescribed accuracy at a prescribed time. The command‐filter technique (CFT) is introduced to eliminate the filtering error, reduce the design complexity, and is used to design the controller consisting of deterministic variables by constructing the appropriate compensation signals. On each switching interval, two scenarios are discussed, and the dynamic ET control strategy based on the relative threshold is designed. By using DSC and the common Lyapunov function (CLF) method, it is demonstrated that all signals in the controlled system are semi‐globally practically finite‐time stable (SGPFTS) in the sense of moment under arbitrary switching. Meanwhile, Zeno behavior is precluded. Finally, numerical simulation results validate the feasibility of the proposed control method.