Higher order fuzzy system identification using subtractive clustering

Article type: Research Article

Authors: Demirli, K.^; | Muthukumaran, P.

Affiliations: Fuzzy Systems Research Laboratory, Department of Mechanical Engineering, Concordia University, 1455 de Maisonneuve Blvd. W., Montreal, Quebec, Canada H3G 1M8

Note: [] Corresponding author. E-mail: demirli@me.concordia.ca.

Abstract: This paper proposes a higher order fuzzy system identification method using subtractive clustering, which is an extended application of subtractive clustering. Minimum error models are obtained through enumerative search of clustering parameters. The results of the enumerative study presented in this paper explain the mechanism behind subtractive clustering and introduce a modification in the penalizing process of subtractive clustering. The results of applying the higher order modeling to both linear and non-linear systems are given in this paper. The comparison with the results of other models shows improvement in the modeling performance by subtractive clustering with higher order system identification technique. The higher order identification method resulted in fewer rules compared to lower order models. Results of case studies on systems of different complexities are also presented.

Journal: Journal of Intelligent and Fuzzy Systems, vol. 9, no. 3-4, pp. 129-158, 2000