southampton

Module overview

The main problem in system identification is deriving mathematical models of dynamical systems (for transfer function, state-space) from data. Such problem arises for example in control, when the complexity of a model or lack of physical insight prevent the development of a model from first principles to be used in designing a controller.

The algorithms to solve the system identification problem rely, for the linear case, on mathematical methodologies formalised with (relatively) simple linear algebra.

The objective of this course is to give a broad but non superficial introduction to some of the main themes in system identification of discrete-time systems: the use of Hankel matrices and of regression and algebraic structures based on time shifts to derive transfer-function and state-space models.

Linked modules

Pre-requisite: ELEC2220 or ELEC2226

Aims and Objectives

Learning Outcomes

Subject Specific Intellectual and Research Skills

Having successfully completed this module you will be able to:

Master and apply basic and advanced techniques of system identification
Use the most efficient models (e.g. state-space, ARMA) applied in system identification, control, and statistics
Master and apply linear-algebraic techniques to compute mathematical models from data

Knowledge and Understanding

Having successfully completed this module, you will be able to demonstrate knowledge and understanding of:

The range of choices available for modelling dynamical systems depending on the physical properties of the system and the available data
How data and their structure can be used to derive mathematical models for dynamical systems

Science and Mathematics

Having successfully completed this module you will be able to:

Master and apply various linear algebraic techniques to compute mathematical models from data

Learning and Teaching

Teaching and learning methods

Lectures, coursework

Study time
Type	Hours
Wider reading or practice	55
Assessment tasks	13
Revision	10
Lecture	36
Preparation for scheduled sessions	18
Follow-up work	18
Total study time	150

Resources & Reading list

Internet Resources

Paul M.J. Van den Hof, System Identification - Data-Driven Modelling of Dynamic Systems, lecture notes, University of Eindhoven..

Textbooks

Peter van Overschee, B.L. de Moor (1996). Subspace identification for linear systems. Springer New York, NY.

Lennart Ljung. System Identification - Theory For the User. Prentice-Hall, Upper Saddle River, New Jersey.

Assessment

Assessment strategy

Coursework and final written examination

Summative

This is how we’ll formally assess what you have learned in this module.

Breakdown
Method	Percentage contribution
Examination	80%
Coursework	20%

Repeat Information

Repeat type: Internal & External

From Data to Dynamical Model: System Identification