Finite-Time Adaptive Neural Prescribed Performance Control for High-Order Nonlinearly Parameterized Switched Systems With Unmodeled Dynamics and Input Quantization

This study focuses on the adaptive prescribed-time neural control for a class of high-order switched systems with nonlinear parameterization in presence of unmodeled dynamics and quantized input. Different from the existing results on finite-time control on basis of adding a power integrator technique, the controller construction and stability analysis are simplified, and the tracking error remains within a set range over any prescribed time. Under the frame of backstepping design, a state feedback controller is designed. During the controller design procedure, Radial basis function (RBF) neural networks with minimal learning parameters are employed to identify the unknown compounded nonlinear functions, and the control input is quantized. Based on Lyapunov stability theory, the closed-loop system’s signals are all assured to be semi-globally uniformly bounded (SGUB), and the tracking error is kept inside a prescribed zone at a finite time. Finally, a numerical simulation is provided to demonstrate the viability and efficacy of the control strategy.