Research

My main research interests concern data-driven modelling for control and predictive maintenance of nonlinear dynamical systems, with a recent focus for robotics and aerospace applications.

Data-driven system identification is a machine learning-based approach for modeling dynamical system behavior by building mathematical models on the basis of experimental input/output data. 

In the past few years, I have been employing more and more regularly deep learning techniques in conjunction with classical system identification and control engineering tools. ​

Current or recent applications include:

  • AI-based RUL prediction. Application to a business jet's air distribution system (collaboration with Dassault Aviation - France)
    • M. Hervé de Beaulieu, M. Jha, H. Garnier, F. Cerbah. Remaining Useful Life prediction based on physics-informed data augmentation​Reliability Engineering and System Safety, December 2024. ​​https://doi.org/10.1016/j.ress.2024.110451​
  • Inertia matrix parameter estimation for satellite attitude control (collaboration with CNES - the French Space Agency)
    • C. Nainer, H. Garnier, M. Gilson, H. Evain, C. Pittet, Parameter Estimation of a Gyroless Micro-Satellite from Telemetry Data, Control Engineering Practice, 123, 105-134, June 2022.  https://doi.org/10.1016/j.conengprac.2022.105134
  • Modelling and control of wireless power transfer systems (collaboration with Wuhan University - China)
    • F. Chen, H. Garnier, Q. Deng, M.K. Kazimierczuk, X. Zhuan, Control-oriented modeling of wireless power transfer systems with phase-shift control, IEEE Transactions on Power Electronics, 35 (2), 2119-2134, February 2020. https://doi.org/10.1109/TPEL.2019.2920863
  • Design of benchmarks for system identification (collaboration with ONERA, France and University of Linköping, Sweden)
    • V. Pascu, H. Garnier, L. Ljung, A. Janot, Benchmark problems for continuous-time model identification: design aspects, results and perspectives, Automatica, 107, 511-517, Sept. 2019. https://doi.org/10.1016/j.automatica.2019.06.011
  • System identification in Robotics (collaboration with University of Veracruz - Mexico)
    • ​​A. Sanchez-Garcia, H. Rios-Figueroa, H. Garnier, G. Quintana-Carapia, E. Rechy-Ramirez, A. Marin-Hernandez, Time-to-contact forecasting based on probabilistic segmentation and system identification, Advanced Robotics, 32(8), 426-442, 2018. http://dx.doi.org/10.1080/01691864.2018.1455604
  • Aerodynamic coefficient identification of a space vehicle (collaboration with Institut Franco-Allemand de Saint-Louis, France)
    • M. Albisser, S. Dobre, C. Berner, M. Thomassin, H. Garnier, Aerodynamic coefficient identification of a space vehicle from multiple free flight tests, Journal of Spacecraft and Rockets, 54 (2), 426-435, March 2017. http://dx.doi.org/10.2514/1.A33587
  • Parameter estimation in battery management systems (collaboration with UC San Diego - USA)
    • B. Xia, X. Zhao, R. de Callafon, H. Garnier, T. Nguyen, C. Mi, Accurate Lithium-ion battery parameter estimation with continuous-time system identification methods.  Applied Energy, 179, 426-436, July 2016. http://dx.doi.org/10.1016/j.apenergy.2016.07.005
  • Parameter estimation for biomedical systems (collaboration with University of Liege - Belgium)
    • J.-B. Tylcz, T. Bastogne, H. Benachour, D. Bechet, E. Bullinger, H. Garnier, M. Barberi-Heyob, A Model-based Method of Characterizing the Pharmacokinetics of Engineered Nanoparticles in Pilot Studies.  IEEE Transactions on NanoBioscience, 14(4), 368-377, June 2015. http://dx.doi.org/10.1109/TNB.2015.2418792