Redundant machine learning architectures for functional safety relevant applications – an evaluation

Doran, Hans Dermot; Ielpo, Gianluca; Ganz, David; Zapke, Michael

Publication type:	Conference poster
Type of review:	Peer review (abstract)
Title:	Redundant machine learning architectures for functional safety relevant applications – an evaluation
Authors:	Doran, Hans Dermot Ielpo, Gianluca Ganz, David Zapke, Michael
et. al:	No
Conference details:	14. Embedded Computing Conference, Winterthur (online), 1. Juni 2021
Issue Date:	Jun-2021
Language:	English
Subjects:	Functional safety; Redundant system; Embedded system; Computer architecture; Machine learning; Neural network
Subject (DDC):	006: Special computer methods
Abstract:	We compare two similar machine learning implementations in a 2oo2 redundant configuration on two platforms, FPGA and GPU, in both an architectural and a performance sense. We examine the real-time characteristics in a theoretical and experimental sense, presenting measurements. We also examine the coordination/synchronisation issues between the redundant components and clearly enumerate the dependability considerations that need to be taken into account in this area. From these considerations we derive the requirements of the synchronisation and voting mechanisms and present first suggestions for these. We critically evaluate the platforms before ending the paper, aimed at decision makers as well as R&D interested, with a review and suggestions for further work/consideration.
URI:	https://digitalcollection.zhaw.ch/handle/11475/23981
Fulltext version:	Published version
License (according to publishing contract):	Licence according to publishing contract
Departement:	School of Engineering
Organisational Unit:	Institute of Embedded Systems (InES)
Appears in collections:	Publikationen School of Engineering

Files in This Item:

There are no files associated with this item.

Show full item record

Doran, H. D., Ielpo, G., Ganz, D., & Zapke, M. (2021, June). Redundant machine learning architectures for functional safety relevant applications – an evaluation. 14. Embedded Computing Conference, Winterthur (Online), 1. Juni 2021.

Doran, H.D. et al. (2021) ‘Redundant machine learning architectures for functional safety relevant applications – an evaluation’, in 14. Embedded Computing Conference, Winterthur (online), 1. Juni 2021.

H. D. Doran, G. Ielpo, D. Ganz, and M. Zapke, “Redundant machine learning architectures for functional safety relevant applications – an evaluation,” in 14. Embedded Computing Conference, Winterthur (online), 1. Juni 2021, Jun. 2021.

DORAN, Hans Dermot, Gianluca IELPO, David GANZ und Michael ZAPKE, 2021. Redundant machine learning architectures for functional safety relevant applications – an evaluation. In: 14. Embedded Computing Conference, Winterthur (online), 1. Juni 2021. Conference poster. Juni 2021

Doran, Hans Dermot, Gianluca Ielpo, David Ganz, and Michael Zapke. 2021. “Redundant Machine Learning Architectures for Functional Safety Relevant Applications – an Evaluation.” Conference poster. In 14. Embedded Computing Conference, Winterthur (Online), 1. Juni 2021.

Doran, Hans Dermot, et al. “Redundant Machine Learning Architectures for Functional Safety Relevant Applications – an Evaluation.” 14. Embedded Computing Conference, Winterthur (Online), 1. Juni 2021, 2021.