FNonlinear System Identification based on Extended Fuzzy Logic

One of the main parts of fuzzy logic theory is fuzzy systems. The extensive research in this area shows the applicability of the systems. However, fuzzy systems carry some shortcomings; especially, when the insignificant parts of the membership functions in the if-section of rules are not empty. This paper based on extended fuzzy logic, which recently introduce by Zadeh, extends fuzzy systems by f-transforming fuzzy inference and introduces an extended fuzzy system. The proposed extended fuzzy system involves the operations of input processing containing fuzzification and validification, output processing containing set-conversion and defuzzification, as well as knowledge processing and making inference. The main novelty of the paper is formulating approximate reasoning within extended fuzzy logic framework. Then the extended fuzzy system as universal approximator is proposed and applied to the identification of nonlinear systems that need more consideration than linear systems. The simulation results show priority of the proposed system when compared to the fuzzy system.