|
04342 Stochastic Adaptive Control
| |
Danish title: Stokastisk adaptiv regulering
|
| Type: Å, Language: EED |
| |
Credit points:
5 point |
|
| | | Previous course: C0414 |
| Offered by:
Department of Mathematical Modelling
(IMM) |
| No credit points with: C0414 |
| Prerequisite: (04041/04040).(50300/72131/36231) |
| Desirable: 04244/50310 |
| Recommended semester:
7th - 9th semester |
| Examination:
Written exam
(13 point scale
) |
| Remarks: In this course control theory is combined with optimization and statistics. It is based on a introductory course in statistics (04041/04040) and an introductory coorse in control (50300/72131/36231). The course is related to 50300, 50320 and 04231. |
| Contact person: |
Niels Kjølstad Poulsen, IMM, Building 321, Tel. +45 4525 3356 |
|
Aim: To give the students a knowledge of methods for:
-Describing and controlling dynamic systems, which are influenced by stochastic disturbances.
-Identification of stochastic systems (i.e. estimation of unknown parameters) and, finally
-Adaptive control of stochastic systems (i.e. simultaneous identification and control). |
Contents: The course consists of three parts. In the first part the stochastic system is assumed to be known and the basic stochastic control theory is given. In order to considering control of an unknown system the second part of the course deals with modelling stochastic systems (i.e. system identification and parameter estimation). In the third and final part of the course the methods described in the previous parts are combined in adaptive control, i.e. control of unknown dynamic systems. The methods are given for discrete time, state space - and transfer function models.
1. Basic stochastic control theory. Stochastic processes and disturbances, Stochastic dynamic systems. Filter theory, especially methods for state estimation, Kalman filter, smooting and prediction. Stochastic control, optimal stochastic control, especially LQG - control, minimal variance control, generalized minimal variance control. Stochastic poleplacement control.
2. Systemidentification. LS -, maximumlikelihood and Bayes estimation of system parameters. Prediction error - and pseudolinearregression methods. Modelvalidation and estimation of structure. Identifiability and identification in closed loop. Persistenly excitation and probing signals. Recursive estimation, convergence analysis and identification of time varying systems.
3. Adaptive Control. The basic self tuner. Explicit and implicit self tuners. Adaptive generalized minimal variance and poleplacement control. The cautions and the optimal adaptive controllers. Convergence analysis. |
|
|
