Controlador bayesiano difuso de velocidad para un motor DC de imán permanente

Abstract

Fuzzy controllers are of great importance in the industry since they are used for different areas of knowledge such as DC motor control. For this reason, the need to know and develop new control strategies that meet the most specific parameters is born. This paper presents the use of Bayesian networks for the improvement of fuzzy rules in fuzzy controllers using data obtained from a small plant, in this case, a permanent magnet DC motor, and using the fundamental characteristics such as error and error variance to establish a new algorithm to improve its behavior. With the use of different Matlab tools and performance indicators such as IAE, ISE, ITAE, and ITSE based on the error analysis, the efficiency of the controller is estimated based on Bayes' Theorem. Also using statistical analysis such as the Wilcoxon test, the new controller is adjusted and its efficiency is determined compared to a conventional Fuzzy controller, showing innovative and interesting results.