Self learning fuzzy controllers using iterative learning tuner

Abstract

This paper describes the design of an adaptive fuzzy controller using iterative learning to tune input membership functions and scaling factor(s). The control scheme consists of a fuzzy controller and learning control laws. People's perception about the meaning of a linguistic variable differs from person to person or even from expert to expert. This difference in perception usually leads to different fuzzy control designs. Some where within these designs lays the required design which meets a specific performance criterion. This paper proposes an approach to tackle this uncertainty in perception, to find the required design using membership function modification. The membership function is adaptively adjusted through iterative learning technique. The results show that the scheme is robust, cost effective and very simple to implement. It makes use of the nonlinearity inherent in the fuzzy systems. This scheme can be used to design fuzzy controllers for different plants by finding the right membership functions to ensure the required design specifications. Designing fuzzy controllers with desired performance specifications is not a trivial task. Even the specification of linguistic variables, key concept in fuzzy system design, can be different from different experts. This scheme tries to fill this gap. Adaptive fuzzy techniques are computationally heavy to implement. The proposed scheme lays out a unique adaptive procedure for designing fuzzy controllers through iterative learning process.