Time-varying non-linear system identification using dynamical adaptive fuzzy systems

• M. Cerrada Universidad de Los Andes-Venezuela
• J. Aguilar Universidad de Los Andes-Venezuela
• E. Colina Universidad de Los Andes-Venezuela
• A. Titli Laboratoire d”™Analysis et d”™Architecture de Syst í¨mes (LAAS-CNRS)-Francia

Abstract

In processes control, it is required to have a model that describes the process behaviour. Identification techniques provide an approximate representation of the real system behaviour, by using a linguistic or mathematical expression, or an algorithm. In this work, a dynamical adaptive fuzzy model is used in order to propose the input-output identification model of a time-varying non-linear system. This class of fuzzy models permit a generic adaptive approach of the process behaviour through the adaptation of the membership functions to the universes of discourse of the model variables.