Presentation 2019-03-26
High Reliability and Safety Protection Method for Applying Artificial Intelligence to Embedded Systems
Yasuhiro Omori, Akihiko Higuchi, Daisuke Kawakami,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) This paper proposes an architecture for embedded systems, which employs machine learning results. The proposed architecture ensures high-reliability and safe operations, while realizing low-cost implementation. Since, machine learning can not guarantee its results, in general it is not easy to apply it to systems under safety requirements. Moreover, since high performance computing units are necessary to carry out machine learning, it is difficult to apply the fault detection and isolation techniques with redundancy to low-cost embedded systems. Our architecture realizes high-reliability by the mechanism that judges the safety of the system from machine learning results, and duplicates the judgement part of the system. In this paper, we show the evaluation results of a prototype of the architecture to demonstrate that our architecture can be applied to ensuring the safety of the system at low cost.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) High Reliability / Functional Safety / Redundancy / Embedded System / Fail Safe / Fail Operational / AI
Paper # SSS2018-32
Date of Issue 2019-03-19 (SSS)

Conference Information
Committee SSS
Conference Date 2019/3/26(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair Masaomi Kimura(Shibaura Inst. of Tech.)
Vice Chair Makoto Ito(Tsukuba Univ.)
Secretary Makoto Ito(NPO RDA)
Assistant Koh Kawashima(Oriental Motor) / Sei Takahashi(Nihon Univ.)

Paper Information
Registration To Technical Committee on Safety
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) High Reliability and Safety Protection Method for Applying Artificial Intelligence to Embedded Systems
Sub Title (in English)
Keyword(1) High Reliability
Keyword(2) Functional Safety
Keyword(3) Redundancy
Keyword(4) Embedded System
Keyword(5) Fail Safe
Keyword(6) Fail Operational
Keyword(7) AI
1st Author's Name Yasuhiro Omori
1st Author's Affiliation Mitsubishi Electric Corporation(Mitsubishi Electric Co.)
2nd Author's Name Akihiko Higuchi
2nd Author's Affiliation Mitsubishi Electric Corporation(Mitsubishi Electric Co.)
3rd Author's Name Daisuke Kawakami
3rd Author's Affiliation Mitsubishi Electric Corporation(Mitsubishi Electric Co.)
Date 2019-03-26
Paper # SSS2018-32
Volume (vol) vol.118
Number (no) SSS-518
Page pp.pp.3-6(SSS),
#Pages 4
Date of Issue 2019-03-19 (SSS)