Parameters Extraction for Fuzzy Modeling of Nonlinear System

Abstract

The modeling and identification of nonlinear systems are important but challenging problems. Because of numerous advantages fuzzy models are often preferred to describe such systems. However, in many cases the generated models are very complex. In the paper, a new fuzzy modeling method of nonlinear system is proposed. The fuzzy model is identified as black-box model with input-output training data. A modified self-organizing map (MSOM) network is developed for generating parameters of fuzzy model. Based on the MSOM, fuzzy rules are determined automatically according to the distribution of training data in the input-output space. Simulating example indicates that the fuzzy modeling method is effective.