ROBUSTNESS AGAINST SEU OF AN ARTIFICIAL NEURAL-NETWORK SPACE APPLICATION

Authors
ASSOUM A RADI NE VELAZCO R ELIE F ECOFFET R
Citation
A. Assoum et al., ROBUSTNESS AGAINST SEU OF AN ARTIFICIAL NEURAL-NETWORK SPACE APPLICATION, IEEE transactions on nuclear science, 43(3), 1996, pp. 973-978
Citations number
14
Categorie Soggetti
Nuclear Sciences & Tecnology","Engineering, Eletrical & Electronic
Journal title
IEEE transactions on nuclear science → ACNP
ISSN journal
00189499
Volume
43
Issue
3
Year of publication
1996
Part
1
Pages
973 - 978
Database
ISI
SICI code
0018-9499(1996)43:3<973:RASOAA>2.0.ZU;2-3
Abstract
We study the sensitivity of Artificial Neural Networks (ANN) to Single Event Upsets (SEU). A neural network designed to detect electronic an d protonic whistlers has been implemented using a dedicated VLSI circu it: the LNeuro neural processor. Results of both SEU software simulati ons and heavy ion tests point out the fault tolerance properties of AN N hardware implementations.