Performance of Fuzzy and Neuro-Fuzzy Controllers on an Unstable Non-Linear Plant

This article presents the results obtained in the performance of a fuzzy controller and a neuro-fuzzy controller in the simulation phase, applied to the control of a non-linear and inherently unstable plant such as the ball and platform device. From the mathematical model of the system, and by using the Matlab® computational tools, both controllers were designed using the classic Mamdani structure and a fuzzy inference system based on adaptive network, or ANFIS, respectively. The verification procedure was carried out by means of the application of input signals in order to print circular and rectangular trajectories on the platform to the ball. The comparison of the responses obtained by simulations shows advantages in favor of the neuro-fuzzy controller, even considering the presence of noise generated in the sensors and the occurrence of possible external disturbances.