Module Description

Linear and Nonlinear System Identifikation

Courses:

Module Responsibility:

Prof. Herbert Werner

Admission Requirements:

None

Recommended Previous Knowledge:

• Classical control (frequency response, root locus)
• State space methods
• Discrete-time systems
• Linear algebra, singular value decomposition
• Basic knowledge about stochastic processes

Educational Objectives:

Professional Competence

Theoretical Knowledge

• Students can explain the general framework of the prediction error method and its application to a variety of linear and nonlinear model structures
• They can explain how multilayer perceptron networks are used to model nonlinear dynamics
• They can explain how an approximate predictive control scheme can be based on neural network models
• They can explain the idea of subspace identification and its relation to Kalman realisation theory

Capabilities

• Students are capable of applying the predicition error method to the experimental identification of linear and nonlinear models for dynamic systems
• They are capable of implementing a nonlinear predictive control scheme based on a neural network model
• They are capable of applying subspace algorithms to the experimental identification of linear models for dynamic systems
• They can do the above using standard software tools (including the Matlab System Identification Toolbox)

Personal Competence

Social Competence

Students can work in mixed groups on specific problems to arrive at joint solutions. 

Autonomy

Students are able to find required information in sources provided (lecture notes, literature, software documentation) and use it to solve given problems. 

ECTS-Credit Points Module:

3 ECTS

Examination:

Oral exam

Workload in Hours:

Independent Study Time: 62, Study Time in Lecture: 28

Course: Linear and Nonlinear System Identification (Lecture)