Modeling Of Nonlinear System Based On T-S Fuzzy Model

Nonlinear systems exist in the world extensively, so it is important to investigate the identification of nonlinear systems. As for nonlinear systems, it is difficult to get the accurate mathematic model. Even we can build mathematic model for those systems, it may be too complicated to design a controller with conventional means.Takagi- Sugeno fuzzy model illustrates the local-rule for every local area with a linear equation and achieves global nonlinearity based on local linearity by fuzzy inference. Considering the merits above of the T-S fuzzy model, this dissertation closely surrounds fuzzy modeling and identification methods for nonlinear systems to discuss and to research.For T-S's system identification, it is possible to separate the premise identify-cation from the consequence identification using fuzzy cluster that can simplify the calculation. Firstly, input variables are selected by the method of stepwise, the variables of much infection to output variables are reserved, the variables of little infection to output variables are eliminated, the complexity of model is properly simplified at the premiss of precision. Secondly, with the result, the input space is clustered by fuzzy c-means clustering algorithm; we get the membership function of the antecedent fuzzy sets. Thirdly, the consequent parameters of rules are calculated by the least squares estimate is presented, the premise and consequence parameters of T-S fuzzy model are optimized by genetic algorithm for advancing its precision. At last, the effectiveness of proposed algorithms are demonstrated by stimulation results of the well-known Box-Jenkins data set.