|
ACI-42076 System Identification, 5 cr |
Matti Vilkko
Lecture times and places | Target group recommended to | |
Implementation 1 |
3.-n. vuosikurssi
Automaatio-, kone- ja materiaalitekniikan tiedekunta Automaatiotekniikan koulutusohjelma Jatko-opiskelijat KV-opiskelijat |
A Computer Examination with Matlab and System Identification Toolbox.
-
The objective is to provide basic theoretical and practical knowledge to student so that he can understand and apply the most important system identification methods.
Content | Core content | Complementary knowledge | Specialist knowledge |
1. | Models and systems - impulse responce - frequency responce - spectrum - linear time invariant systems, model structures - noise model | Periodogram, windowing | Blackman-Tukey, Hamming, etc. |
2. | Methods: - non-parametric time and frequency domain methods - parameter estimation methods, residual, LR, LSE - parameter computation, iterative methods - recursive estimation | Prewhitening, ETFE, RLS algorithm | QR decomposition, Newton, Quasi-Newton, LM, gradient computation, matrix inversion lemma |
3. | Application of identification: - experiment desing - data preprocessing - identification objective - model structure selection - validation | identification of feedback system | Integrator in target system |
4. | Identification tool usage: - process of identification - Matlab System Identification Toolbox - model comparison in time and frequency domain, residual analysis | detection of delay in the system |
Type | Name | Author | ISBN | URL | Edition, availability, ... | Examination material | Language |
Book | System Identification, Theory for the User | Lennart Ljung | 0-13-656695-2 | ISBN 0-13-656695-2 | Suomi |
Course | O/R |
ACI-20090 Mallinnus ja simulointi | Recommended |
MAT-20500 Todennäköisyyslaskenta | Recommended |
Course | Corresponds course | Description |
|
|
Description | Methods of instruction | Implementation | |
Implementation 1 | Lectures Excercises |
Contact teaching: 0 % Distance learning: 0 % Self-directed learning: 0 % |