Logo ČVUT
CZECH TECHNICAL UNIVERSITY IN PRAGUE
STUDY PLANS
2023/2024
UPOZORNĚNÍ: Jsou dostupné studijní plány pro následující akademický rok.

Estimation, filtering and fault detection

The course is not on the list Without time-table
Code Completion Credits Range Language
XE35EFF Z,ZK 7 3P+1S English
Garant předmětu:
Lecturer:
Tutor:
Supervisor:
Department of Control Engineering
Synopsis:

The course is devoted to classical and modern methods for filtering of measured signals, corrupted by random noise or systematic errors, estimation of parameters of dynamic systems (ARX/ARMAX models), and stochastic state-space models estimation and filtering (Kalman filter). Futher topics include model-based fault detection and isolation, applications of alternative or parallel models based on the Bayesian approach, reliable numerical algorithms and software for estimation, filtering and fault detection. Applications of presented concepts will be further applied for specific aerospace problems - guidance and navigation, merging of complementary measurements, estimation of satellite position etc.

Requirements:

Basic course on linear algebra: solving linear systems, basic matrix decompositions (LU, Cholesky, QR, SVD), eigenvectors/eigenvalues, singular values, conditioning. Basic course on probability and statistics. Basic knowledge of stochastic signals and systems (correlation function, covariance matrix, white and colored noise definition and properties).

Syllabus of lectures:

1.Introduction to estimation, filtering and fault detection (EFFD).

2.Classical filtering (Butterworth filters etc.). Review of probability and random signals.

3.Filtering, prediction and smoothing. Classical and probabilistic filtering.

4.Dynamical systems and models. Stochastic linear state-space system.

5.Linear system and stochastic input.

6.Estimation of random variables. Estimation criteria, optimal estimators.

7.Kalman filter.

8.Kalman filter II.

9.Kalman filter, parallel and alternative models.

10.Convergence and stability of Kalman filter. Frequency properties.

11.Stochastical methods of identification. ARX, ARMAX, OEM models.

12.Stochastical methods of identification. Batch and on-line identification. 13.Fault detection and isolation. Problem formulation, case studies.

14.Model-based fault detection and isolation methods. Links to parameter estimation and state space filtering.

Syllabus of tutorials:

1.Introduction to estimation, filtering and fault detection (EFFD).

2.Butterworth filters etc. Review of probability and random signals.

3.Filtering, prediction and smoothing. Classical and probabilistic filtering.

4.Dynamical systems and models. Stochastic linear state-space system.

5.Linear system and stochastic input.

6.Estimation of random variables. Estimation criteria, optimal estimators.

7.Kalman filter.

8.Kalman filter II.

9.Kalman filter, parallel and alternative models.

10.Convergence and stability of Kalman filter. Frequency properties.

11.Stochastical methods of identification. ARX, ARMAX, OEM models.

12.Stochastical methods of identification. Batch and on-line identification. 13.Fault detection and isolation. Problem formulation, case studies.

14.Model-based fault detection and isolation methods. Links to parameter estimation and state space filtering.

Study Objective:
Study materials:

1.Chin-Fang Lin: Modern Navigation, Guidance and Control Processing, Prentice Hall, 1991.

2.Brian D.O. Anderson, John B. Moore: Optimal Filtering, Dover Publications, 2005.

Note:
Further information:
https://moodle.dce.fel.cvut.cz/course/view.php?id=14
No time-table has been prepared for this course
The course is a part of the following study plans:
Data valid to 2024-03-27
Aktualizace výše uvedených informací naleznete na adrese https://bilakniha.cvut.cz/en/predmet12366704.html