Digital Gaussian membership function circuit for neuro-fuzzy hardware

Abstract

A digital design of a Gaussian membership function circuit is presented. It allows positioning of the function on the universe of discourse to be selected, and its width or variance to be adjusted. The circuit is therefore suitable for implementing digital neuro-fuzzy systems with learning capabilities. The circuit allows control of the delay-time/accuracy rate, and adds a new parameter to the adjustable parameter set of the system that may be used to improve the learning algorithm and optimise system architecture.