FUZZY SYSTEM AS A PARAMETER ESTIMATOR OF NONLINEAR DYNAMIC FUNCTIONS

Abstract

In this paper Adaptive Fuzzy Systems are used as intelligent identification systems for nonlinear plants. A new technique to design the fuzzy system that relies on the minimisation of a loss function is presented. The design technique uses the centres of the fuzzy sets (labels) at the antecedent part of the Rule Base as the estimated parameters. This parametrisation has the Linear in The Parameters (LITPs) characteristic that allows standard parameter estimation technique to be used to estimate the parameters of the fuzzy system. The combination of the fuzzy system and the estimation method then performs as a nonlinear estimator. If several fuzzy sets are defined for the input variables at the antecedent part, the fuzzy system (“fuzzy estimator”) then behaves as a collection of nonlinear estimators where different rule's regions have different parameters. The proposed scheme is potentially capable of estimating the parameters of highly nonlinear plants. Simulation examples, which use plants with highly nonlinear gain, show the power of the proposed estimation scheme in comparison to estimation using linear model.