Sensors & Transducers



Vol. 249, Issue 2, February 2021, pp. 54-64





1, * Yannick FOURASTIER, 2 Claude BARON, 3 Hakima CHAOUCHI,
4 Carsten THOMAS, 5 Thomas BOURGEAU, 6 Jean-Paul SMET,
7 Viktor YEHOROV and 8 Philippe ESTEBAN



1 CodEUrope SAS, 244 Route de Seysses, 3100 Toulouse, France

2 Université Toulouse, INSA, LAAS-CNRS, ISAE-Supaéro, 7 Av. Colonel Roche, 31200 France

3 Institut Polytechnique de Paris, Telecom SudParis, 9, rue Charles Fourier, 91011, Evry, France

4 HTW Berlin, Treskowalee 8, 10318, Berlin, Germany

5 KB Digital, 1291 Commugny, Switzerland

6 Nexedi GmbH, Technology Centre Munich, Agnes-Pockels-Bogen 1, 80992 München, Germany

7 University of Odessa, Lab Mironaft, Kanatna St, 112-114, 65000 Odessa, Ukraine

8 Université de Toulouse III, LAAS-CNRS, 7 Av. Colonel Roche,
31200 Toulouse, France

E-mail: yannick.fourastier@codeurope.eu



Received: 7 October 2020 /Accepted: 30 November 2020 /Published: 28 February 2021





Abstract: Complex industrial systems are increasingly software-defined, rapidly evolving into autonomous, self-adaptive processing at industrial fields. They are constituting worlds of things, wherein the internet of things is heavily developing with local semantic features. Also, artificial intelligence technologies are spreading fast in the industrial context, improving the operations efficiency. However, Edge computing systems involving artificial intelligence must also continuously ensure reliable and safe operations. This paper assesses the state of the art of cognitive technologies with their relevance for artificial intelligence implementation in industrial cyber-physical systems. It then introduces a state of research on artificial intelligence safety assurance practices. Implementation through three use cases at the software stack is illustrated with a virtual operating system that operates the data space in the industrial layout. More precisely, the industrial development is illustrated with a real case of swarm computing for advanced robotics powered with SlapOS. Although still at an early stage, it shows experimentally how Edge operating systems are evolving fast as a kind of AI intensive software.


Keywords: Artificial intelligence, Cyber-physical system, Industrial edge, Swarm processing, Virtual operating system, Digital twin, Safety of intended functionality.

___________________________________________________________________