System Identification (ETF AEO IS 4860) |
|
General information |
|
Module title | System Identification |
Module code | ETF AEO IS 4860 |
Study | ETF-B |
Department | Control and Electronics |
Year | 1 |
Semester | 2 |
Module type | Mandatory |
ECTS | 7 |
Hours | 60 |
Lectures | 39 |
Exercises | 21 |
Tutorials | 0 |
Module goal - Knowledge and skill to be achieved by students |
|
| Course objective is to give students skills and knowledge referring to concepts and methods used in analysis and identification of processes and systems through analytical, grapho-analytical and experimental methods of system identification. Within the course, through examples of various types of processes and using software tools, students will learn how to define optimum forms and identify the parameters of the mathematical models that best describe given series of experimental data, validating the model on the same series of data in which it was developed. <br> |
|
Syllabus |
|
| 1. Identification methods and procedures. Elementary definitions and classification. ARX models and linear least square methods. Model groups, model structures, identification and formal aspects. Identification of model structures. Identification of multivariable black-box model's structure. Time variable and nonlinear systems models. Black-box nonlinear models. Neural networks, wavelets and classical models. Fuzzy models. Nonparametric methods of identification in time and frequency domains. <br> 2. Estimation and identification with parametric methods. Identification through techniques of regression and recursive least square methods. Identification methods with instrumental variables. Recursive methods with error prediction. Identification methods using random signals. Identification through sequential learning. Automation classification algorithms in identification techniques. <br> 3. Experiment design. Closed-loop identification. Sampling interval and presampling factors selection. Data pre-processing. Identificational criterion selection. Norm selection: robustness. Model structure selection and model validation. Real system identification: illustration of IDENT toolbox and SIT package in MATLAB framework and working with SIT package. <br> |
|
Literature |
|
| Recommended | 1. Lecture notes and slides (will be available at the Web site) <br> 2. Lennart Ljung: System identification- Theory for the user, 2nd ed. Prentice Hall, 1999 <br> 3. The Mathworks, MATLAB software package. <br> 4. The Mathworks, IDENT software package. |
| Additional | |
Didactic methods |
|
| Lectures objective is to give detailed review of all teaching modules of the course. Lectures are presented in an auditory in a way to enable students to easily follow its dynamics and to immediately observe concepts and methods that seem vague. After presenting every segment of the course, the teacher demonstrates examples and tasks which enable students to learn terminology, work with instruments and methodologies shown during the lectures. Additional examples and examination tasks are considered and solved in laboratory (under assistant-tutor's guidance). These activities are structured in such way that level of students - knowledge and skills, meant to be achieved in this course, is being continuously examined through laboratory hours and seminar tasks. <br> |
|
Exams |
|
| Through the course, student gains points according to following system. <br> Attendance to lectures and laboratories: 10 points, student which misses lectures and/or laboratories more than three times cannot get points for these activities. <br> Two project tasks (seminars) equally allocated through semester: maximum 50 points. <br> Student which in the end of the course has less than 20 points has to take the course again. <br> Student which in the end of the course has 40 or more points can take final exam; this exam is consisted of discussion on candidate's seminar tasks, or other student's seminar tasks, and answers to questions referring to course subjects. <br> Final verbal examination is worth maximum 40 points. <br> Student which has gained more than 20 but less than 40 points during the course takes corrective exam. Corrective exam is structured in the following way: <br> - written examination, structured in the same way as seminar task; on this examination student solves tasks from subjects he/she did not pass in seminar task. <br> - verbal examination, structured in the same way as final verbal exam. <br> Student can take verbal corrective examination only if after passing written corrective examination has made total score of 40 or more points; this score is made of points gained through: attendance, seminar tasks and passing written corrective examination. <br> Verbal corrective examination is worth maximum 40 points. <br> |
|
Aditional notes |
|