## Data Driven Modelling of Linear and Nonlinear Systems (2016)

Welcome to Data Driven Modelling of Linear and Nonlinear Systems

Description:

Often models for dynamic systems are impossible, very difficult or just too time consuming to establish based on first principles. Occasionally model structures can be formulated but the values of the parameters are not known.  Often, as for many control applications, a simple approximate model is more useful compared to a complicated model developed from first principles.  In all of the above situations it is necessary to provide methods for estimating the parameters and perhaps the whole model structure must be identified and all the parameters must be estimated based on available time series.  Methods for this is within engineering called system identification'', while within statistics the term "system modelling'' is used.

Basic approaches for system modelling includes prediction error methods for estimating parameters in linear SISO ARMAX models based on open loop data.  However, in many situations the system or the data calls for a more complex framework for modelling, like closed loop data, MIMO, time-varying or nonlinear models.

The purpose of this PhD course is to give the participant a comprehensive knowledge on both basic and more advanced aspects of systems modelling. The goal is to enable the students with knowledge and tools for stochastic modelling of physical systems.  The participant should be able to apply software for parameter estimation for the model structures.  The software used are IDENT from MATLAB and CTSM-R - Continuous Time Stochastic Modelling for R developed at DTU COMPUTE.

Prerequisites:

Knowledge on methods for describing dynamical systems in continuous and discrete time, and on statistics and stochastic processes at master level.

Organizer: Professor Rafal Wisniewski, e-mail: raf@es.aau.dk
Lecturers: Associate Professor Torben Knudsen & Professor Henrik Madsen

ECTS: 3.0

Time: June 6-10, 2016

Place:

Zip code:

City:

Number of seats: 50