Orthogonal optimized-choice algorithm for non-linear systems identification based on fuzzy model

In the paper, the structure determination and parameter estimation for the non-linear systems are presented by means of the dynamic fuzzy model. The parameters estimation of fuzzy model is independent of each other by means of the orthogonal method. The most significant fuzzy rules are selected into the fuzzy model based on the “Innovation-Contribution” criterion and some other information criteria. The orthogonal method which is the stepwise-regression algorithm with appending rules or deleting rules has nothing to do with the selected term sequence of fuzzy rules. The simulation example is studied to demonstrate the effectiveness of the proposed algorithm.