Fuzzy rule-based neural appointed-time control for uncertain nonlinear systems with aperiodic samplings

Abstract

In this paper, a fuzzy rule-based neural appointed-time control scheme for uncertain nonlinear systems with aperiodic samplings is proposed. To guarantee tracking properties with preassigned convergence time and remove the high dependency on accurate initial states of system, we construct an improved prescribed performance control (IPPC) scheme on the basis of a hyperbolic cosecant performance function with a finite-time behavior. In light of the unavailability of partial states and uncertainties, by combining a fuzzy wavelet neural network (FWNN) with a state observer, a FWNN-based state observer is developed, which can simultaneously approximate unavailable system states and unknown lumped disturbances with a remarkable accuracy. In addition, aimed at eliminating the problem of parameter updating explosion caused by overlarge learning dimensions, a minimum-learning-parameter (MLP) technique is embedded in the FWNN-based state observer, where the norm of weight matrix is employed for online adaptive updating. Furthermore, an event-triggered control scheme with relative thresholds is synthesized within the framework of dynamic surface control (DSC), which can allow for aperiodic samplings to save communication and actuating resources. Meanwhile, a Nussbaum type function is introduced to solve the issue of unknown control coefficients. The significant features of our work are twofold: (1) Appointed-time tracking performances with much fewer sampling times are assured. (2) An enhanced robustness against uncertainties is achieved with a lower computational complexity and the requirement of full-state measurements is relaxed. Finally, two simulation examples are performed to validate the effectiveness and advantages of the proposed control scheme.