Simultaneous design of membership functions and rule sets for type-2 fuzzy controllers using genetic algorithms

Abstract

A Type-2 Fuzzy logic controller adapted with genetic algorithm, called type-2 genetic fuzzy logic controller (T2GFLC), is presented in this paper to handle uncertainty with dynamic optimal learning. Genetic algorithm is employed to simultaneous design of type-2 membership functions and rule sets for type-2 fuzzy logic controllers. Traditional fuzzy logic controllers (FLCs), often termed as type-1 fuzzy logic systems using type-1 fuzzy sets, cannot handle large amount of uncertainties present in many real environments. Therefore, recently type-2 FLC has been proposed. The type-2 FLC can be considered as a collection of different embedded type-1 FLCs. However, the current design process of type-2 FLC is not automatic and relies on human experts. The purpose of our study is to make the design process automatic. The evolved type-2 FLCs can deal with large amount of uncertainties and exhibit better performance for the mobile robot. Furthermore, it has outperformed their type-1 counterparts as well as the traditionally designed type-2 FLCs.