Title :
Nonlinear system identification based on modified ANFIS
Author :
José Kleiton Ewerton da Costa Martins;Fábio Meneghetti Ugulino de Araújo
Author_Institution :
Universidade Federal do Rio Grande do Norte, Departamento de Engenharia da Computaç
fDate :
7/1/2015 12:00:00 AM
Abstract :
This article aims to present the nonlinear system identification by the method of modified ANFIS. The modified ANFIS is a structure proposed that is based on the traditional structure of ANFIS with some modifications as shown in the article. The importance of the choice of method parameters and its influence on the system will be discussed. For this, the identification of a coupled system of tanks with nonlinear dynamics is performed. System identification will be performed by changing the inputs and order of the consequent model and then will perform a review of the systems. The results confirm the simplicity of modified ANFIS in comparison with the traditional ANFIS while have good performance in the identification of nonlinear systems.
Keywords :
"Training","Biological neural networks","Computational modeling","Fuzzy systems","Software","Nonlinear dynamical systems"
Conference_Titel :
Informatics in Control, Automation and Robotics (ICINCO), 2015 12th International Conference on