Adaptive Neural Network Tracking of a Class of Switched Nonlinear Systems with Time-varying Output Constraints

Abstract

An approximation-based adaptive design problem for output-constrained tracking of a class of switched pure-feedback nonlinear systems is investigated under arbitrary switchings. All switched nonlinearities are assumed to be unknown. Contrary to the existing control results for uncertain switched pure-feedback nonlinear systems where the number of the used function approximators should be equal to the order of the systems, an adaptive control scheme based on only two neural networks is designed by using a system transformation and the common Lyapunov function method, regardless of the order of the system. In the proposed controller, the output constraints are used to establish designable time-varying bounds on the tracking performance. The stability and the constraint satisfaction of the resulting closed-loop system are shown in the sense of Lyapunov stability criterion. Finally, simulation examples are provided to illustrate the effectiveness of the proposed methodology.