Data-driven Modeling of Dynamic Systems

Up-to-date information will be distributed via Moodle only!

Please register for the lecture via the TUCaN system of TU Darmstadt. If you have problems accessing Moodle, please contact

General information about the lecture

Lecturer Dr.-Ing. Eric Lenz
Scope WS (2+1)
Prerequisites Systemdynamik und Regelungstechnik I or Optimal and Predictive Control or Control of Distributed Cyber-Physical Systems or Model Predictive Control and Machine Learning
Desirable prerequisites Fundamentals of statistical signal theory
Examination written, oral

Depending on the number of exam registrations, the exam will be either written or oral. The type of examination will be announced shortly after the end of the registration period.
In the summer semesters, the examination is expected to be written.
Old exams Old exams can be downloaded via Moodle.
Supplementary and advanced courses Modeling, Simulation and Optimization
Recommended literature Isermann, Münchhof: Identification of Dynamic Systems, Springer Verlag, 2011.
Ljung, L.: System Identification: Theory for the user. Prentice Hall information and systems sciences series. Prentice Hall PTR, Upper Saddle River NJ, 2. edition, 1999.
Pintelon, R.; Schoukens, J.: System Identification: A Frequency Domain Approach. IEEE Press, New York, 2001.

Content of the lecture

Aim of this course

Introduction to the determination of mathematical models for the behavior of dynamic systems from measured signals.

Topics

  • Introduction
  • Basics
    • Stochastics
    • Signal processing (Fourier series, Fourier transform, DTFT, DTF, …)
    • Stochastic and deterministic signals
  • Non-parametric identification
    • Frequency response estimation with periodic and time-limited signals
  • Parametric identification
    • Determination of parameters
    • Identification via the minimization of the output error
    • Identification using time and frequency domain data with the linear least squares method
      • Numerical methods
      • Instrumental variable method
    • Identification using Kalman filters

Exam winter semester 2024/25

Exam oral
Date 03.04.2025
Time Please register for a time slot by 14.02.2025 via the link published in the Moodle forum of this lecture (2024)!
Room S3|10-406