Subtractive fuzzy C-means clustering approach with applications to fuzzy predictive control

摘要: To identify T-S models, this paper presents a so-called "subtractive fuzzy C-means clustering" approach, in which the results of subtractive clustering are applied to initialize centers and number rules order perform adaptive clustering. This method not only regulates division inference system input output space determines relative member function parameters, but also overcomes impacts initial values on performance. Additionally, orthogonal least square algorithm is employed parameters consequents linearize systems over every sample time, ultimately resulting entire models. With approach available, model predictive control established, along with corresponding algorithms derived, as well simulations carried out explicitly demonstrate effectiveness proposed method.