FUZZY IDENTIFICATION METHOD BASED ON A NEW OBJECTIVE FUNCTION

Many fuzzy models on modeling and controlmethods have been developed since Takagi-Sugeno model was proposed1] .The main prob-lems of the fuzzy identifying method based on T-S model appear in the following two aspects:　　 1 It is important what heuristic knowledgeis chosen to acquire the premise structure of theT- S model.This problem is related to the opti-mal partition of input space of premise part andmodeling accuracy.　　 2 Itis crucial whatidentifying method is u-tilized to estimate …

FUZZY IDENTIFICATION METHOD BASED ON A NEW OBJECTIVE FUNCTION

A method of fuzzy identification based on a new objective function is proposed. The method could deal with the issue that input variables of a system have an effect on the input space while output variables of the system do not exert an influence on the input space in the proposed objective functions of fuzzy clustering. The method could simultaneously solve the problems about structure identification and parameter estimation; thus it makes the fuzzy model become optimal. Simulation example demonstrates that the method could identify non-linear systems and obviously improve modeling accuracy.