Optimization of fuzzy logic implementation

Abstract

An architecture for implementing fuzzy-logic inference, together with the tools to optimally synthesize fuzzy logic circuits under this architecture, is proposed. The algorithms for finding the parameters for this architecture are presented. The author discusses how, computer-aided-design tools can be built to help fuzzy logic designers to explore the design space. In particular, the design process is divided into two phases, tuning the membership functions and synthesizing hardware implementations. The former is application-dependent, and previously no well-established automatic methods had been developed. A flexible, interactive design environment that allows designers to easily modify the plant model, the inference mechanism, and the membership functions, and quickly find out what the result is will be highly desirable.<>