EE 531 - System Identification

Announcement (Semester 1, 2022)

• All course information are officially in CourseVille. Homework assignments, data files, or other online resources are in CourseVille.
• You can use this page to download course handouts only.

Book

System Identification Book by Jitkomut Songsiri

• The book is the main reference of this course (written in Thai). Students can purchase the book from the catalog link in the book’s page. It is available in both color-printed version and e-book.

Lecture videos

Videos playlist on YouTube: System identification: parametric approach

Lecture notes

1. Introduction
2. Review on linear system and random processes
3. Review on linear algebra
4. Model parametrization
5. Input signals
6. Linear least-squares
7. Significance tests for LS
8. Variations on least-squares
9. Instrument variable methods
10. Prediction error methods
11. Statistical estimation
12. Subspace methods
13. Model selection and model validation

the topics not taught currently

0. Transient and Frequency analysis
1. Correlation analysis
2. Fourier analysis
3. Spectral analysis
4. Identification of closed loop systems
5. Some practical aspects
6. Recursive identification methods

Term Project

Each year we aim to assign the problems of fitting models to real-world data as group assignments. Students learn a general description of the problem before the midterm and the project starts after the midterm. Students develop the problem statement, estimation formulation, experimental results and write the report throughout the six weeks until the final week. Each group writes a progress report and present it weekly. Students should apply the estimation techniques learned from class or more advanced/alternative methods from research literature. The full reports from the previous years include:

Project 2021

1. Electrical load forecasting by SARIMAX model

   • Contributor: Phing Lim

2. EEG-based valence and arousal estimation using wavelet scalogram and convolutional networks

   • Contributor: Napat Samsow and Totok Nugroho

Project 2018

0. Project description and weekly details (2018)

1. Learning Brain Network Differences from EEG Data

   • Contributor: Parinthorn Manomaisaowapak

2. Photovaltaic System Modeling

   • Contributor: Parinthorn Manomaisaowapak
   • Contributors: Kitinan Boonravee and Jeerapat Jitnuant

3. Time Series Models of Stock Price

   • Contributor: Parinthorn Manomaisaowapak
   • Contributor: Maytus Piriyajitakonkij

Project 2017

0. Project description and weekly details (2017)

1. Modeling of Photovoltaic System

   • Contributors: Janenarong Klomklao

2. An Identification of Building Temperature System

   • Contributors: Chanthawit Anuntasethakul/ Natthapol Techaphanngam/ Natdanai Sontrapornpol
   • an extension of this project was published in ECTI-CON paper, 2018

3. An EEG subspace identification

   • Contributors: Satayu Chunnawong

4. Recursive Estimation of Solar Irradiance using Time-Series Model

   • Contributors: Tony Fang
   • the project was not finished completely

Project 2016

0. Project description and weekly details (2016)

1. Estimation of Autoregressive with Exogeneous Inputs Model for fMRI Data

   • Contributors: Nattaporn Plub-in/ Patawee Prakrankamanant/ Nop Polboon/ Ranyaphat Hongpipatsak/ Morokot Cheat

2. Solar Forecasting using Time Series Models

   • Contributors: Maxime Facquet/ Supachai Suksamosorn/ Veenakorn Suphatsatienkul/ Vichaya Layanun

Project 2015

0. Project description and weekly details (2015)

1. A Convex Formulation of Structural Equation Modeling (SEM) in fMRI Study

   • Contributors: Anupon Pruttiakaravanich/ Auangkun Rangsikunpum/ Pusit Suriyavejwongs/ Tawan Luprasong

2. Parameter estimation of Gumbel distribution for flood peak data

   • Contributors: Piyatida Hoisungwan (co-advisor from Dept. of Water Resource)/ Jitin Khemwong/ Tiwat Boonyawiwat/ Tanakorn Kriengkomol

3. Rainfall Grid Interpolation from Rain Gauge and Rainfall Predicted from Satellite Data

   • Contributors: Piyatida Hoisungwan (co-advisor from Dept. of Water Resource)/ Petchakrit Pinyopawasutthi/ Pongsorn Keadtipod/ Tanut Aranchayanont

4. Fitting Vector Auto Regressive Model to Electroencephalography (EEG) Signals

   • Contributors: Nuntanut Raksasri/ Akasit Aupaiboon/ Pawarisson Jaitahan

MATLAB Data Files

• Least-squares fitting: data-LScostheta.mat
• State trajectory estimation: data-state-traj-est.mat
• Estimation of initial conditions: data-state-traj-est.mat
• Least-squarse fitting: data-ls-fitting.mat
• Estimation of a first-order system: data-est1storder.mat
• Estimation of scalar AR proceses: nikkei_feb11_feb12.mat
• Estimation of stable vector AR processes: data-vec-ar.mat and data-vec-ar-short.mat
• MAP estimation of a linear model: data-map-linmodel
• Recursive least-squares with a forgetting factor: data-rls-ff
• Comparison between recursive least-squares and recursive instrument variable method: data-rls-riv
• MAP estimation of an ARX model: data-arxmap
• Simulation and One-Step Prediction: data-predicted-simulated.mat
• Topology selection in Gaussian graphical model: data-gaussian-gm
• Choosing scalar AR model order: data-modelsel-ar
• Choosing the effective length of FIR models: data-modelsel-FIR
• Consistency of IV and LS estimates: data-consistent-iv3
• Optimal Nonlinear estimator: data-nonlinest1
• Optimal Nonlinear estimator: data-nonlinest
• Robust least-squares with Laplacian uncertainty: data-robust-LS-laplacian
• Compare the bias and variance of sparse LS estimate: data-cmp-var-bias-sparseLS
• Estimation of time series trends: data-ls-season.mat
• Estimation of two models that share a parameter: data-varls-twomodels-commoncoef.mat
• Estimation of a first-order state-space model: data-ls-1stss-knowx.mat

Course Information

Lectures:EE404, Mon Wed 8:00-9:30 am Textbooks:

• 10. Songsiri, System Identification, 2021 (in Thai).
• 12. Ljung, System Identification: Theory for the User, 2nd edition, Prentice Hall, 1999
• 20. Soderstrom and P. Stoica, System Identification, Prentice Hall. 1989 (available for download from the author website)
• 7. James and D. Witten, T. Hastie, and R. Tibshirani, An Introduction to Statistical Learning with Applications in R, Springer, 2013
• 20. Hastie, R. Tibshirani, and J. Friedman, The Elements of Statistical Learning, Springer, 2009
• J.P. Norton, An Introduction to Identification, Dover, 1986
• P.Van Overschee and B.D. Moor, Subspace identification for linear systems: Theory—Implementation—Applications, Kluwer Academic Publishers, 2012
• 13. Verhaegen and V. Verdult, Filtering and System Identification: A Least-square Approach, Cambridge University Press, 2007.

Related textbooks:  

• P.S.P. Cowpertwait and A.V. Metcalfe, Introductory Time Series with R, Springer, 2009
• 18. 8. Shumway and D.S. Stoffer, Time Series Analysis and Its Applicattions: with R Examples, Springer, 2009
• R.A. Horn and C.R. Johnson, Matrix Analysis, 2nd edition, Cambridge, 2013

Grading:Refer to what has been announced in My CourseVille