Deep neural networks-based output-dependent intermittent control for a class of uncertain nonlinear systems

Uncertain dynamics is prevalent in nonlinear control systems, and neural networks (NNs) is a conventional method commonly employed to address uncertainty in control systems. In contrast to previous approaches, deep neural networks (DNNs) demonstrates superior capabilities in enhancing the performance of unknown function approximation in nonlinear systems. This paper investigates the use of DNNs-based output information intermittent control for nonstrict feedback nonlinear systems (NFNS). Initially, the semiglobal practical finite-time stability (SGPFTS) framework is proposed for uncertain intermittent control systems, along with the introduction of an output-information intermittent control scheme. Subsequently, leveraging the established finite-time intermittent control result, DNNs, and Lyapunov stability theory, sufficient conditions are presented to ensure the attainment of SGPFTS for NFNS. Furthermore, the theoretical findings are applied to practical scenarios, and corresponding simulation results are presented to demonstrate the effectiveness of the theoretical outcomes and validate the practical viability of the proposed control method.