Program of the 2nd Spring Doctoral School on

Data-driven Model Identification of Dynamical Systems

Polytech Nancy, France

March, 26-29, 2018

 

The 4-days Spring Doctoral School will consist of a series of lectures followed by computer exercises.

For the computer exercices, it is required that participants bring their own laptop with the latest version of Matlab installed with stand alone license.

March 26, 2018 - 8h30-9h00
Registration - Hall of Building F - Polytech Nancy

    March 26, 2018 - 9h00-17h00
    Lecturer: Prof. Johan Schoukens (Vrije Universiteit Universiteit, Belgium)

      One-day introductory course on System Identification

       

      March 27-28, 2018 - 9h00-17h00
      Lecturer: DR. Xavier Bombois (Ecole centrale de Lyon, France)

      A- two day course on data-driven approaches for discrete-time model identification

      Overview

       

        • Theme 1: Introduction;concepts; discrete-time signal and system analysis.
        • Theme 2: Parametric (prediction error) identification methods: prediction criterion and model structures, linear and pseudo-linear regressions, conditions on data, statistical and asymptotic properties.
        • Exercise: Hands on identification of discrete-time parametric models.

         

        • Theme 3: Parametric (prediction error) identification methods: model set selection and model validation, approximate modelling.
        • Theme 4: Non-parametric identification (ETFE); experiment design.
        • Computer exercise : Getting hands on experience with the Matlab System Identification toolbox.

        March 29, 2018 - 9h00-12h00
        Lecturer:  Prof. Hugues Garnier and Prof. Marion Gilson (University of Lorraine, France)

        An overview of data-driven approaches for direct continuous-time model identification

          • Theme: Overview of continuous-time linear model identification. Basic SVF-based and optimal instrumental variable (IV)-based estimators. Benefits for practical applications. Software aspects.
          • Computer exercise. Getting hands on experience with the CONTSID toolbox methods with real data.

          March 29, 2018 - 14h00-16h00
          Lecturer: Mathieu Cuenant (MathWorks)

          Identification and control - An industrial perspective