Defense of thesis "Modeling of an artificial intelligence-based decision support for cyber-physical systems: Application to maintenance in the transportation field"".

French summary

Cyber-physical systems (CPS) are complex, interconnected systems, where digital technologies converge with the physical world, having revolutionized many business sectors. The Internet of Things (IOT) has also encouraged the massive collection of data, contributing to the development of "Big Data". The development of the IOT and PCS has, however, brought with it challenges, mainly linked to massive data management and decision-making. The development of approaches such as Data Analytics and Machine Learning, however, offers solutions by identifying "hidden" patterns or features within this massive data.

The present dissertation addresses this decision-making issue using Artificial Intelligence approaches and techniques, with a particular focus on the Transport application domain. Modern transportation systems, considered as complex cyber-physical systems, integrate a wide variety of mechatronic equipment. They are increasingly autonomous, equipped with sensors that enable them to perceive their environment, and with the means to interact with fleet managers. A global approach to fleet management can improve various functions such as maintenance planning and operations management.

The thesis focuses on cybernetic loops associated with the operation of cyberphysical systems, with an emphasis on the maintenance activity. It proposes decision-making assistance by characterizing cybernetic loops, specifying the needs of decision-makers, and developing Machine Learning approaches while respecting the requirements of genericity and technological independence. The modeling of this assistance is based on a holonic decomposition of the SCP, and uses Rasmussen's typology at the decision-making processing level. A methodological guide accompanies this approach.

The concepts proposed in this thesis work have been validated through two industrial collaborations. A first collaboration with the Moroccan company STMF optimized the maintenance of a fleet of trucks transporting hazardous materials. A second collaboration with ALSTOM aims to improve the reliability growth of railway rolling stock by detecting early warning signals of failure.

Keywords

Decision support, Cyber-physical system, Cyber loops, Artificial intelligence, Maintenance.

Jury composition

Rapporteurs:

Mr Philippe THOMAS, Maître de Conférences, HDR, Université de Lorraine
M. Hassan EL GHAZI, Professor, Institut National des postes et télécommunications de Rabat

Examiners:

Mrs. Hasna CHAMLAL, Habilitated Professor, FSAC, Université Hassan II de Casablanca
Mme. Malika ZAZI, Professeure, ENSAM-Rabat, Université Mohammed V de Rabat
Mr. Olivier SENECHAL, Professor, Université Polytechnique Hauts-de-France
Mr. Maroua NOUIRI, Professor, ENSAM-Rabat Ms Maroua NOUIRI, Senior Lecturer, University of Nantes

Ms.

Co-Thesis supervisors:

Mr Yves SALLEZ, Professor, Université Polytechnique Hauts-de-France
Mr. Badr ABOU EL MAJD, Professor, FSR, Université Mohammed V, Rabat