An efficient analog Hamming distance comparator realized with a unipolar memristor array: a showcase of physical computing

We propose and demonstrate a novel physical computing paradigm based on an engineered unipolar memristor that exhibits symmetric SET switching with respect to voltage polarity. A one-dimensional array of these devices was sufficient to demonstrate an efficient Hamming distance comparator for two str...

Full description

Saved in:
Bibliographic Details
Main Authors: Ge, Ning, Yoon, Jung Ho, Hu, Miao, Merced-Grafals, E. J., Davila, Noraica, Strachan, John Paul, Li, Zhiyong, Holder, Helen, Xia, Qiangfei, Williams, R. Stanley, Zhou, Xing, Yang, J. Joshua
Other Authors: School of Electrical and Electronic Engineering
Format: Article
Language:English
Published: 2017
Subjects:
Online Access:https://hdl.handle.net/10356/83449
http://hdl.handle.net/10220/42646
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
Description
Summary:We propose and demonstrate a novel physical computing paradigm based on an engineered unipolar memristor that exhibits symmetric SET switching with respect to voltage polarity. A one-dimensional array of these devices was sufficient to demonstrate an efficient Hamming distance comparator for two strings of analog states represented by voltages from the physical world. The comparator first simultaneously applies the two sets of voltages to the array of memristors, each of which is initially in its high resistance state and switches to its low resistance state only if the two voltages applied on that memristor differ by more than the switching threshold. An accurate analog representation of the Hamming distance is then obtained by applying a reading voltage to the memristors and summing all the resultant currents. The comparator with a small footprint can directly process analog signals and store computation results without power, representing a promising application for analog computing based on memristor crossbar arrays.