Research

Current Research interests:
Safe Reinforcement Learning
Physics Informed Neural Networks
Deep Learning based prognostics.

PhD Thesis Co-direction (Present): 

Soha Kanso (2021-...),
Safe Reinforcement Learning for dynamical systems under degradation "  See  Here . Funded by CRAN, Université de Lorraine.

Theo Rutschke (2023-...),  
Physics-driven identification of nonlinear systems for reinforcement learning , See Here. Funded by CRAN, Université de Lorraine.

PhD Thesis Co-direction (Past): 

Martin Herve de Beaulieu (2020-2023)
Subject: "Identification and prognosis of state of health of non-linear systems through deep learning. Application to predictive maintenance of business aircraft " (2020-2023), In collaboration with Dassault Aviation. 
PhD Thesis

Post Doctoral Research Co-supervision
Julien Thuillier, (2021-2023), Subject: Health Aware Control Design of Liquid Propulsion Rocket Engine.
Funded by CNES. See papers here,here,here..

Industrial Projects (On-going)

  1. Co-PI (Co-principle Investigator) , Research project and co-director of Doctoral thesis, in collaboration with Dassault Aviation.
    Subject: "Identification and prognosis of state of health of non-linear systems through deep learning. Application to predictive maintenance of business aircraft " 
     
  2. Co-PI (Co-principle Investigator)  Research Projects with The French National Centre for Space Studies (CNES)
    (Kick-off, September 2023) Design of Learning approaches for Health aware control of Reusable liquid rocket engine.

Industrial Projects (Past)

  1. Co-PI (Co-principle Investigator)  Research Projects with The French National Centre for Space Studies (CNES)

 

                                  

International Projects:
Our project proposal with Chili in collaboration with Prof. Hugues Garnier, has been selected by the ECOS-Sud Programme.
The collaboration starts in 2022 for 3 years. The goal is to develop deep learning-based methods for identifying continuous-time nonlinear systems.

National level Projects:
Leader of French national Level Action Project: Health Aware Control Design in Dynamic Systems is a GDR MACS action (see here)

Master Thesis Supervision (MSc)

Mohammad Chelouati, " Conception d’un système de loi de commande fondée sur l’état de santé du procédé consacrée aux propulseurs réutilisables ".
See the scientific paper here.
Soha Kanso, " Estimation de la durée de vie résiduelle d’une chambre de combustion LRE " . See the scientific paper here.
George Claudiu ANDREI, " Deep Reinforcement Learning for Dynamical Systems ". See Report and scientific paper.

 

Active Reviewer Scientific Journals:

 

 

Guest Editor of Special Issue on "Intelligent Systems for Fault Diagnosis and Prognosis", Sensors. See here.