Publication Details

Triple Modular Redundancy Used in Field Programmable Neural Networks

KRČMA Martin, LOJDA Jakub and KOTÁSEK Zdeněk. Triple Modular Redundancy Used in Field Programmable Neural Networks. In: Proceedings of IEEE East-West Design & Test Symposium. Novi Sad: IEEE Computer Society, 2017, pp. 1-6. ISBN 978-1-5386-3299-4.

Czech title

TMR technika použitá na FPNN

Type

conference paper

Language

english

Authors

Krčma Martin, Ing. (DCSY FIT BUT)
Lojda Jakub, Ing., Ph.D. (DCSY FIT BUT)
Kotásek Zdeněk, doc. Ing., CSc. (DCSY FIT BUT)

Keywords

Tripple Modular Redundancy, FPNN, Aritificial Neural networks, Fault Tolerance

Abstract

This paper presents the concepts of FPNA and FPNN, used for the approximation of artificial neural networks in FPGAs and discusses the usage of TMR technique in order to reach a fault tolerance. The schemes of the FPGA implementation are presented. The results of experiments determining the FPGA resources utilization with different usage of the TMR technique are provided.

Published

2017

Pages

1-6

Proceedings

Proceedings of IEEE East-West Design & Test Symposium

Conference

15th IEEE EAST-WEST DESIGN & TEST SYMPOSIUM, Dr Zorana Đinđića 1, 21101, Novi Sad, RS

ISBN

978-1-5386-3299-4

Publisher

IEEE Computer Society

Place

Novi Sad, RS

DOI

10.1109/EWDTS.2017.8110128

UT WoS

000426878200101

EID Scopus

2-s2.0-85040520044

BibTeX

@INPROCEEDINGS{FITPUB11446,
   author = "Martin Kr\v{c}ma and Jakub Lojda and Zden\v{e}k Kot\'{a}sek",
   title = "Triple Modular Redundancy Used in Field Programmable Neural Networks",
   pages = "1--6",
   booktitle = "Proceedings of IEEE East-West Design \& Test Symposium",
   year = 2017,
   location = "Novi Sad, RS",
   publisher = "IEEE Computer Society",
   ISBN = "978-1-5386-3299-4",
   doi = "10.1109/EWDTS.2017.8110128",
   language = "english",
   url = "https://www.fit.vut.cz/research/publication/11446"
}