Contribution à la commande adaptative floue pour les systèmes non linéaires discrets incertains

Abstract

The aim of this thesis is to develop adaptive fuzzy control structures combined with different approaches such as the optimal approach, backstepping and synergetic approach for a class of non-linear discrete systems SISO uncertain and perturbed. In this perspective, three robust fuzzy control structures based on the above techniques have been developed. Fuzzy systems are used to approximate the dynamics of nonlinear systems where some of their parameters are adapted online to ensure the stability of the closed loop system via Lyapunov synthesis. The first technique uses direct dynamic heuristic programming based on the filtered tracking error for the optimal tracking problem while minimizing an "objective" function. The second approach uses the backstepping technique based on the Minimum Learning Parameter (MLP) technique to solve the problem of the explosion of complexity inherent to the backstepping technique. The third approach based on synergetic control, because of its robustness to external uncertainties and disturbances while ensuring constant switching frequency operation. These three control structures have been validated, by simulation under the Matlab environment, through three non-linear discrete systems SISO