Open Access System for Information Sharing

Login Library

 

Article
Cited 52 time in webofscience Cited 58 time in scopus
Metadata Downloads

Neuromorphic Character Recognition System With Two PCMO Memristors as a Synapse SCIE SCOPUS

Title
Neuromorphic Character Recognition System With Two PCMO Memristors as a Synapse
Authors
Sheri, AMHwang, HJeon, MLee, BG
Date Issued
2014-06
Publisher
Institute of Electrical and Electronics Engineers
Abstract
Using memristor devices as synaptic connections has been suggested with different neural architectures in the literature. Most of the published works focus on simulating some plasticity mechanism for changing memristor conductance. This paper presents a neural architecture of a character recognition neural system using Al/Pr0.7Ca0.3MnO3 (PCMO) memristors. The PCMO memristor has an inhomogeneous barrier at the aluminum and PCMO interface which gives rise to an asymmetrical behavior when moving from high resistance to low resistance and vice versa. This paper details the design and simulations for solving this asymmetrical memristor behavior. Also, a general memory read/write framework is used to describe the running and plasticity of neural systems. The proposed neural system can be produced in hardware using a small 1 K crossbar memristor grid and CMOS neural nodes as presented in the simulation results.
URI
https://oasis.postech.ac.kr/handle/2014.oak/13590
DOI
10.1109/TIE.2013.2275966
ISSN
0278-0046
Article Type
Article
Citation
IEEE Transactions on Industrial Electronics, vol. 61, no. 6, page. 2933 - 2941, 2014-06
Files in This Item:

qr_code

  • mendeley

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

Related Researcher

Researcher

황현상HWANG, HYUNSANG
Dept of Materials Science & Enginrg
Read more

Views & Downloads

Browse