Piotr Andrzej Kowalski, Szymon Łukasik, Małgorzata charytanowicz, Piotr Kulczycki Data-Driven Fuzzy Modeling and Control with Kernel Density Based Clustering Technique Polish Journal of Environmental Studies, vol. 17, nr 4C, ss. 83-87, 2008
Streszczenie:
Deriving parameters and structure of fuzzy model for a dynamical system by means of a clustering procedure is a very popular and frequently applied technique in fuzzy identification. The aim of the paper is to present a novel method of fuzzy model formulation based on this approach. Introduced algorithm is based on clustering method employing nonparametric kernel density gradient estimation. Proposed technique is automatic and attains functionality free from arbitrary assumptions concerning "shapes" of data samples, the number of rules and any other user defined parameters. Illustrative results of computer simulations using MATLAB scientific environment are enclosed. The outcome of such experimental verification demonstrates high efficiency of proposed technique in fuzzy controllers synthesis and nonlinear systems modeling.
Słowa kluczowe:
fuzzy modeling, fuzzy control, clustering, kernel density estimation.