Memristor Crossbar-based Hardware Implementation of Fuzzy Membership Functions

TL;DR

This paper presents simple memristor crossbar circuits capable of implementing customizable fuzzy membership functions with various shapes and resolutions, advancing hardware design for fuzzy logic and AI applications.

Contribution

It introduces novel memristor crossbar-based circuits for hardware implementation of flexible fuzzy membership functions, enabling diverse shapes and resolutions.

Findings

Circuits can realize fuzzy functions of arbitrary shapes.

The design supports high-resolution fuzzy membership functions.

Potential use in evolutionary and adaptive systems.

Abstract

In May 1, 2008, researchers at Hewlett Packard (HP) announced the first physical realization of a fundamental circuit element called memristor that attracted so much interest worldwide. This newly found element can easily be combined with crossbar interconnect technology which this new structure has opened a new field in designing configurable or programmable electronic systems. These systems in return can have applications in signal processing and artificial intelligence. In this paper, based on the simple memristor crossbar structure, we propose new and simple circuits for hardware implementation of fuzzy membership functions. In our proposed circuits, these fuzzy membership functions can have any shapes and resolutions. In addition, these circuits can be used as a basis in the construction of evolutionary systems.