Model Identification and Adaptive Systems

• Course program
• Lecture notes
• Projects
• Exams
• Software and teaching material

  ---

  Course program

  The goal of this course is to provide the background for advanced modelling and data analysis techniques. The course has both a theoretical and a practical flavour. This course is the natural extension of the course "Model Identification and Data Analysis". The course focuses on the following topics:

  1. Recursive and adaptive methods for parameter identification: Recursive Least Squares (RLS), Extended Least Squares (ELS), Recursive Maximum Likelihood (RML). Forgetting factor. Adaptive and predictive control: self-tuning control.
  2. Adaptive filtering and active noise control applications: Adaptive digital filtering. Digital filter structures: FIR, IIR. Standard and modified LMS algorithm. The active noise control setting. Broadband feedforward control algorithms: FXLMS, leaky FXLMS, Filtered-U Recursive LMS. Narrowband feedforward control. Periodic disturbance rejection. Disturbance frequency estimation.
  3. State space methods: Estimation, prediction and filtering based on the Kalman Filter. Using the Kalman filter for parameter estimation (gray-box identification). Extended Kalman Filter for non-linear systems. Virtual sensors. Subspace identification methods.
  4. Nonlinear model identification: block-oriented models, NARX models, neural networks, Radial Basis Functions. Model structure selection.

  The final test examination is structured in two parts: a written test and a project.

  ---

  Lecture notes

  0. Course Presentation
  1. Introduction
  2. Recursive Least Squares
  3. Recursive Maximum Likelihood
  4. Adaptive control
  5. Adaptive filtering
  6. Active Noise Control: Introduction and applications
  7. Broadband feedforward ANC: The FxLMS algorithm
  8. Broadband feedforward ANC: The FuLMS algorithm
  9. Narrowband feedforward ANC
  10. Adaptive notch filtering with frequency tracking
  11. On-line secondary path modeling techniques
  12. Kalman filtering: Introduction
  13. Kalman filtering: The steady state predictor
  14. Kalman filtering: Comparison with the Kolmogorov-Wiener approach
  15. Particle filtering
  16. Multivariable model identification
  17. Nonlinear model identification
  18. Identification of polynomial NARX models
  19. Generalized frequency response functions
  20. Artificial neural networks

  Contact me by e-mail to obtain lecture notes not available anymore.

  ---

  Projects

  A (provisional) list of suggested projects is available at this link. Contact me for project assignments and further details. Personal project proposals are most welcome.

  ---

  Written test exams (PDF)

  • Academic Year 2015-16
    • 1st call (11/07/16)
    • 2nd call (25/07/16)
    • 3rd call (14/09/16)
    • 4th call (27/09/16)
    • 5th call (14/02/17)
  • Academic Year 2016-17
    • 1st call (28/06/17)
    • 2nd call (28/07/17)
    • 3rd call (07/09/17)
    • 4th call (25/01/18)
    • 5th call (15/02/18)
  • Academic Year 2017-18
    • 1st call (20/06/18)
    • 2nd call (18/07/18)
    • 3rd call (10/09/18)
    • 4th call (09/01/19)

  ---

  Software e materiale didattico

  • Matlab m-files on recursive/adaptive algorithms
  • Matlab m-files on ANC methods