Fuzzy modelling using Kalman filter

Fuzzy modelling is an important topic in fuzzy sets theory and applications. An efficient method for automatically constructing a Takagi-Sugeno (TS) fuzzy model, where only the input-output data of the identified system are available, is presented. The TS fuzzy model is automatically generated by the process of structure and parameter identification. In the structure identification step, a clustering method based on the Gustafson-Kessel algorithm is proposed. In the parameter identification step, the Kalman filter algorithm is applied twice to choose the parameter values in the premise and consequent parts from the given membership functions defined point-wise and from input-output data. The effectiveness of this approach is demonstrated using two examples.