Controlador nebuloso por modos deslizantes adaptativo

Abstract

In this work, an Adaptive Sliding Neurofuzzy Controller was proposed and its parameters had been established and well characterized. In this way, some concepts of the theories of neurofuzzy networks, sliding mode control and adaptive control had been treated. The considered controller presents only three inference rules and a low computational effort. It was shown, step by step, the conception of the said controller, since the idea of a fuzzy controller without adaptation until its final form with adaptation. All steps were instanced with the sphere levitation system in a magnetic field. The controller stability analyses for nonlinear systems were shown. Aspects of convergence of the adaptation had been also analysed. Simulations and tests had been done with the intention to consolidate the theory and to explore the main characteristics of that controller, such as: inherent robustness, adaptation, immunity the noises and imperfections in the practical applications.