Desarrollo e Implementación del Algoritmo de un Controlador Neuro-Difuso Tipo ANFIS aplicado a un Sistema Aerodinámico TRMS (Twin Rotor MIMO System)

Abstract

The main objective of this research work was to develop the algorithm of programming of a neuro-diffuse system known as ANFIS (Adaptive Network based Fuzzy Inference System) and apply it in modeling and control processes on a TRMS (Twin Rotor MIMO System) aerodynamic system. At the beginning the analysis of its operation and architecture is realized, aiming at being able to develop and implement its programming algorithm using Matlab software and its Simulink environment. The performance of the ANFIS controller was evaluated by implementing it on the TRMS control system together with a conventional PID controller, to do later a comparative statistical analysis using the absolute error integral (IAE) obtained from the outputs of both control systems. Prior to the physical implementation of the ANFIS controller, simulated working tests were done using as plant the mathematical model of the TRMS system.