Advanced System Identification

Course responsible person: Niels Kjølstad Poulsen, Department of Applied MAthematics and Computer Science, build. 324 (phone +45 4525 3356)

Course description: The focus of the course is decided between participants. The course deals with methods for modelling dynamic system based on measurements and system knowledge. First, we shall survey the underlying concepts in system identification, such as parameter estimation methods, consistency and asymptotic distribution of parameters and transferfunction estimate, model validation and structure determination. We shall deal with choices of prefilters and the effect on the identification results. The focus will also be on identifibility and systems operating in closed loop.

Second, the intention is to study the area of identification and control, ie. to study the interplay between identification and control. If time allows we shall have a look at model reduction and system identification based on subspace methods. This part of the course will be based on selected papers.

Course content: Estimation methods, asymtotic properties, reduced order models Identification for control, Control relevant prefiltering, Model reduction and subspace methods.

Institution: Technical University of Denmark

Danish name of institution: Danmarks Tekniske Universitet

Official abbreviation: DTU

Department: Department of Applied Mathematics and Computer Science (Building 324)

Dates: Subject to discussion. Suggestion is 6 days (selected by the participants) during autoum term

Course arrangement: Lectures on selected days combined with exercises and homework

Teaching arrangement: Lectures by participants, exercises

Language: English

Place: Department of Applied Mathematics and Computer Science, Building 324, DTU, DK-2800 Lyngby Number of participation hours: 36

Number of actual working hours: 120

Estimated part of obligatory 1/2 year of course: 14%

The course is part of the following Ph.D.-program: Mathematics, Physics and Informatics

The course is in this program: Optional

Is the course open to other than Ph.D.-students: Yes (Ph.D. students have priority); higher level undergraduates

Evaluation result: Pass/Not-pass

Further informations: on the course can be obtained at the following: E-mail address: nkpo@dtu.dk.