An Approach for Building Takagi-Sugeno Fuzzy Models

An approach for building accurate and fast running T-S models is proposed based on summarization of the experience of nonlinear modeling.Firstly,the number of fuzzy rules is reduced with clustering algorithm in order to improve the running speed of the model,and the number of data in each rule is determined.Then,the parameters of the state matrixes in the rules of T-S model are determined initially with regressive least squares algorithm.Finally,all of the parameters of the T-S model are determined accurately with gradient descent algorithm.The simulation with a real example shows the effectiveness of the approach.