Error injection analysis for triple modular and penta-modular redundancies

Ryo Terada, Minoru Watanabe

Research output: Chapter in Book/Report/Conference proceedingConference contribution

6 Citations (Scopus)

Abstract

Recently, radiation-hardened embedded systems are demanded for robots that must function in high-radiation environments such as those prevailing at the Fukushima Daiichi and Chernobyl nuclear power plants. Recent news reports have described the discovery of a 650 Sv/h radiation region near the containment vessels of the Fukushima Daiichi nuclear power plant. Even under such a dauntingly intense radiation environment, workable robots are required. This paper therefore present a high-soft-error-tolerant 16-bit adder unit based on penta-modular redundancy (SMR) in addition to triple modular redundancy (TMR). The soft-error tolerances of the TMR and SMR were analyzed theoretically. Such TMR and SMR circuits were implemented onto a field programmable gate array (FPGA). Then error-injection experiments for the TMR and SMR were done on the FPGA. Analytical results have revealed that SMR 16-bit adder circuit is useful for heavy radiation environments.

Original languageEnglish
Title of host publication2017 6th International Symposium on Next Generation Electronics, ISNE 2017
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781538630969
DOIs
Publication statusPublished - Jul 5 2017
Externally publishedYes
Event6th International Symposium on Next Generation Electronics, ISNE 2017 - Keelung, Taiwan, Province of China
Duration: May 23 2017May 25 2017

Publication series

Name2017 6th International Symposium on Next Generation Electronics, ISNE 2017

Conference

Conference6th International Symposium on Next Generation Electronics, ISNE 2017
Country/TerritoryTaiwan, Province of China
CityKeelung
Period5/23/175/25/17

ASJC Scopus subject areas

  • Hardware and Architecture
  • Electrical and Electronic Engineering

Fingerprint

Dive into the research topics of 'Error injection analysis for triple modular and penta-modular redundancies'. Together they form a unique fingerprint.

Cite this