Identificazione dei Modelli ed Analisi dei Dati

Modulo di 56 ore per il III anno della Laurea in Ingegneria dell'Automazione e per la Specialistica in Ingegneria dell'Automazione (N.O.) al Dipartimento di Ingegneria dell'Università di Ferrara. A.A. 2006/2007.

Course Programme

Lecture 1. Introduction - General introduction to modelling and system identification. (i) Theory and experiment based modelling methods; (ii) Parametric and non-parametric models and identification methods; (iii) Procedure of system identification. Reading: L. Ljung, From Data to Model: A Guided Tour of System Identification, Linköping University, Sweden, Report No. LiTH-ISY-R-1652, 1994. Lecture Notes: Lecture 1. (Lesson Slides PDF Format, 4MB); (Lesson Slides PDF Zipped Format, 3.5MB). (Lesson Slides PDF, 2 slides per page, 4MB).

Lecture 2. Non-recursive (off-line) methods. (i) Least-Squares (LS) method and its variants; (ii) Instrumental variable methods; (iii) Prediction error methods. Lecture Notes: Lecture 2. (Lesson Slides PDF Format, 2.5MB); (Lesson Slides PDF Zipped Format, 2MB). (Lesson Slides PDF, 2 slides per page, 2MB).

Lecture 2.5. Nonparametric Identification. Input Signals for Identification. Identification Conditions. Lecture Notes: Lecture 2.5. (Lesson Slides PDF Format, 1.6MB); (Lesson Slides PDF Zipped Format, 1.4MB). (Lesson Slides PDF, 2 slides per page, 1.4MB).

Lecture 3. Recursive (on-line) methods. (i) Recursive Least-Squares (RLS) methods; (ii) Tacking and forgetting factor techniques. Lecture Notes: Lecture 3. (Lesson Slides PDF Format, 1.5MB); (Lesson Slides PDF Zipped Format, 1.2MB). (Lesson Slides PDF, 2 slides per page, 1.2MB).

Lecture 5.5. Remarks on Persistent Excitation Conditions. Practical Aspects. Lecture Notes: Lecture 5.5. (Lesson Slides PDF Format, 0.4MB); (Lesson Slides PDF Zipped Format, 0.38MB). (Lesson Slides PDF, 2 slides per page, 0.38MB).

Lecture 4.6. Model Structure Determination. Model Validation. Theoretic Results and Practical Examples. Lecture Notes: Lecture 4.6. (Lesson Slides PDF Format, 2.6MB); (Lesson Slides PDF Zipped Format, 2.3MB). (Lesson Slides PDF, 2 slides per page, 2.3MB).

Lecture 4. PEM Identification Method. Minimisation Techniques and Numerical Algorithms. Lecture Notes: Lecture 4. (Lesson Slides PDF Format, 1.5MB); (Lesson Slides PDF Zipped Format, 1.2MB). (Lesson Slides PDF, 2 slides per page, 2MB).

Lecture 4.5. Instrumental Variable Method (IVM). Instrument Theory and Numerical Implementation Algorithms. Lecture Notes: Lecture 4.5. (Lesson Slides PDF Format, 1.7MB); (Lesson Slides PDF Zipped Format, 1.5MB). (Lesson Slides PDF, 2 slides per page, 1.6MB).

Lecture 5. Summary and Practical Aspects. An Application Example with the System Identification Toolbox of Matlab. Lecture Notes: Lecture 5. (Lesson Slides PDF Format, 6.6MB); (Lesson Slides PDF Zipped Format, 4.6MB). (Lesson Slides PDF, 2 slides per page, 6.3MB).

Lecture 6. Course Summary on System Identification. Lecture Notes: Lecture 6. (Lesson Slides PDF Format, 2.8MB); (Lesson Slides PDF Zipped Format, 2.2MB). (Lesson Slides PDF, 2 slides per page, 2.6MB).

Lecture 4.8. Closed Loop Identification. Theoretic Results and Practical Examples. Lecture Notes: Lecture 4.8. (Lesson Slides PDF Format, 1.2MB); (Lesson Slides PDF Zipped Format, 1MB). (Lesson Slides PDF, 2 slides per page, 1MB).

Download

Simple example of Matlab function for the identification of a 2nd order ARX model: parameter and estimation error computations (written by Silvio Simani, February 2007): (file in Matlab format, 3KB).
Simple AR and ARX data files (Silvio Simani, February 2007): (text comment file for the ARX example); (file in Matlab dat format). (text comment file for the AR example); (file in Matlab dat format); (file in Matlab format, generation example of the sun spots); (data file in ascii .dat format, time series of the sun spots).
Matlab function "newhank.m" for construction of Hankel matrices (written by Silvio Simani, 2007): (file in Matlab .m format).
Matlab function "struct_selectARX.m" for ARX model structure determination (written by Silvio Simani, 2007): (file in Matlab .m format).
Matlab function "struct_selectARX_val2.m" for ARX model structure determination. Note that this function uses two sets of identification and validation data (written by Silvio Simani, 2007): (file in Matlab .m format).
Matlab functions for testing residual whiteness and Chi-square test (written by Silvio Simani, 2007): (file in Matlab.m format).
Matlab functions for testing residual whiteness and Chi-square test. Note that this function uses two sets of identification and validation data (written by Silvio Simani, 2007): (file in Matlab.m format).
Data file from a simulated SISO (BJ) model (generated by Silvio Simani, 2007): (file in ascii .dat format); (Matlab .m file for BJ model data generation)
Data file from a simulated ARMA model (single output time series, generated by Silvio Simani, 2007): (comment file in ascii .txt format); (data file in ascii .dat format).
Process Data Sequences: DaISy (Database for the Identification of Systems). Developed, maintained, and hosted by SISTA .
General Process Data Sequences for ARX, AR, ARMAX, ARMA and ARIMAX model identification examples: (Data text files in zipped format).
Process data sequences for the Buffer Vessel Example and few Matlab programme files: (Data and Matlab function files in zipped format); (Data file in .mat Matlab format, 49KB); (Description file in .txt format).
Process data sequences for the energised R-phase 400 kV three phase transformer and the Matlab function of the case study: (Data and Matlab function files in zipped format).
Process Data Sequences from SISTA website: (Dynamic Process Data Files in zipped format downloaded from SISTA website).
K. Sigmon, "Matlab Primer Third Edition". University of Florida, Florida, Second Edition ed., 1992. (file in formato pdf ).
SID, "System Identification Toolbox". System Identification Graphical User Interface for Matlab. (file in formato pdf ).

Simani Home Page

Dipartimento di Ingegneria