Time: 12:00-17:30 on Friday from 16th Nov. to 14th Dec. 2012, (12:00-14:30 on 14th Dec.)
Begin: 16th Nov. 2012
Consulting hours: according to agreement in the first course
Examination: oral exam; time will be announced during the lecture.
Registration: An early registration before the lecture is not required. Students who would like to attend the lecture can register in the first lecture.
About the course: This course introduces the fundamentals of soft-computing methods like neural networks, fuzzy logic method, and so forth and their applications in system identification and control. As the counterpart of model-based control, the introduced methods are classified as data-driven techniques because they are fundamentally data-based and require no mathematical system description in advance.
URL der Veranstaltung: http://www.uni-due.de/srs/v-ddt_en
Prerequisite knowledge:- Linear Algebra - Control Techniques (Control Theory preferable)
Recommended reading material: [1]. M. Norgaard, O. Raven, N. K. Poulsen, et al.. Neural Networks for Modeling and Control of Dynamic Systems. London, UK: Springer, 2000 [2]. S. Haykin. Neural Networks: A Comprehensive Foundation. New Jersey, USA: Prentice-Hall, 1999 [3]. K. M. Passino, S. Yurkovich. Fuzzy Control. Menlo Park, USA: Addison-Wesley, 1998 [4]. G. Chen, T. T. Pham. Introduction to fuzzy sets, fuzzy logic, and fuzzy control systems. 2001
Modules of the course:
(NOTE: The scheme is not discretized weekly)
Modules
|
Topics
|
School hour
|
1
|
Introduction and fundamentals
- Definition and classification of data-driven techniques (soft-computing)
- Fundamentals of dynamical systems
- Basics of system identification and control
- Case study: identification and control of linear systems
|
8h
|
2
|
Neural network method Part I: Feedforward Neural Networks (FNN)
- Introduction and Training mechanisms
- System identification and control with FNN
- Case study
|
7h
|
3
|
Neural network method Part II: Recurrent Neural Networks (RNN)
- Definitions and Training mechanisms
- System identification and control with RNN
- Case study: inverted pendulum control
|
5h
|
4
|
Fuzzy logic method
- Introductions to fuzzy logic method
- Fuzzy identification and fuzzy control system
- Combination of neural networks and fuzzy logic
- Case study
|
6h
|
5
|
Support vector machine (SVM)
- Theory and training mechanisms
- Case study: SVM in stability analysis
|
3h
|
6
|
Other methods such as genetic algorithm etc.
|
1h
|
|