Abstract

Alaa Eldin Ahmed Khalil
A Novel Technique for the Design and Real-Time Implementation of Non-Linear Equidistance Fuzzy Set.
Abstract This paper Describes the conditions and steps to design non-linear equidistant membership functions in order to achieve the generalization, smoothing and usage of fuzzy logic in modeling, simulating and controlling both linear and nonlinear systems. Matlab program was used to validate the work in graphical form. The advantages of using the nonlinear membership function are mentioned. The paper also illustrate a new fuzzy modeling algorithm to achieve real-time implementation of fuzzy system based on nonlinear membership functions. Protus simulator is used for the implementation of FLC using the new fuzzy modeling algorithm. The three implemented methods are: the actual exponential function, the Taylor expansion and look-up table. All these techniques were programmed in C-language for all nonlinear membership functions, then downloaded and executed on an AVR microcontroller. Then a comparison is held between various representation techniques in terms of program size and execution time to determine the most efficient and optimal one.