T-S fuzzy model identification based on chaos optimization

Abstract

Nonlinear system identification by fuzzy model has been widely applied in many fields, for fuzzy model has the ability to approximate any nonlinear system to a given accuracy. In this paper, a nonlinear system identification algorithm based on Takagi-Sugeno (T-S) fuzzy model is proposed. Considering that the fuzzy space structure of T-S fuzzy model has great influence upon precision and effect of the final identification result, a new fuzzy clustering method based on chaos optimization, which is believed to be more accurate in data clustering, is adopted to partition the fuzzy space and obtain the fuzzy rules. Based on the fuzzy space partition and the subsequent structure parameters, the least square method is used to identify the conclusion parameters. Typical nonlinear function has been simulated to test the precision and effect of the proposed identification technique, and finally the algorithm has been successfully applied in the boiler-turbine system identification. © 2008 Springer-Verlag Berlin Heidelberg.