CTU FEE Moodle
Estimation, filtering and detection
This is a grouped Moodle course. It consists of several separate courses that share learning materials, assignments, tests etc. Below you can see information about the individual courses that make up this Moodle course.
Estimation, filtering and detection (Main course) A3M35OFD
| Credits | 6 |
| Semesters | Winter |
| Completion | Assessment + Examination |
| Language of teaching | Czech |
| Extent of teaching | 3P+1C |
Annotation
This course will cover description of the uincertainty of hidden variables (parameters and state of a dynamic system) using the probability language and methods for their estimation. Based on bayesian prblem formulation principles of rational behsavour under uncertainty will be analysed and used to develp algorithms for estimation of parameters of ARX models and Kalman filtering including the extensions.
We will demonstrate numerically robust implementation of the algorithms applicable in real life problems for the areas of industrial process control, robotics and avionics. We will extend the methods for linear gaussian systems to a more generic problems using Monte Calro approach. The course will also cover multimodel approach and its use for the fault detection and isolation and introduction to adaptive control.
We will demonstrate numerically robust implementation of the algorithms applicable in real life problems for the areas of industrial process control, robotics and avionics. We will extend the methods for linear gaussian systems to a more generic problems using Monte Calro approach. The course will also cover multimodel approach and its use for the fault detection and isolation and introduction to adaptive control.
Study targets
No data.
Course outlines
1.Problem formulation, estimation methods
2.Bayesian approach to uncertainty description
3.Dynamic system model, probabilistic state definition
4.Identification of ARX model parameters
5.Tracking of time varuing parameters, forgetting, role of prior informaiton.
6.Numerically robust implementaiton for real time parameter tracking
7.Stochastic system, Kalman filter.
8.Kalman filtr for colour noise, extended Kalman filter, adaptive Kalman filter.
9.Stochastic dynamic programming, certainty equivalence principle.
10.Adaptive control, cautious and certainty equivalent strategies, dual control.
11.Probabilistic method for fault detection and isolation
12.Utilizaiton of multiple models
13.Nonlinear estimation, local approximation
14.Global aproximation, Monte Carlo Kalman filter
2.Bayesian approach to uncertainty description
3.Dynamic system model, probabilistic state definition
4.Identification of ARX model parameters
5.Tracking of time varuing parameters, forgetting, role of prior informaiton.
6.Numerically robust implementaiton for real time parameter tracking
7.Stochastic system, Kalman filter.
8.Kalman filtr for colour noise, extended Kalman filter, adaptive Kalman filter.
9.Stochastic dynamic programming, certainty equivalence principle.
10.Adaptive control, cautious and certainty equivalent strategies, dual control.
11.Probabilistic method for fault detection and isolation
12.Utilizaiton of multiple models
13.Nonlinear estimation, local approximation
14.Global aproximation, Monte Carlo Kalman filter
Exercises outlines
Laboratory covers work on individual assignments/projects.
.
.
Literature
No data.
Requirements
No data.
Estimation, filtering and detection AE3M35OFD
| Credits | 6 |
| Semesters | Winter |
| Completion | Assessment + Examination |
| Language of teaching | English |
| Extent of teaching | 3P+1C |
Annotation
This course will cover description of the uincertainty of hidden variables (parameters and state of a dynamic system) using the probability language and methods for their estimation. Based on bayesian prblem formulation principles of rational behsavour under uncertainty will be analysed and used to develp algorithms for estimation of parameters of ARX models and Kalman filtering including the extensions.
We will demonstrate numerically robust implementation of the algorithms applicable in real life problems for the areas of industrial process control, robotics and avionics. We will extend the methods for linear gaussian systems to a more generic problems using Monte Calro approach. The course will also cover multimodel approach and its use for the fault detection and isolation and introduction to adaptive control.
We will demonstrate numerically robust implementation of the algorithms applicable in real life problems for the areas of industrial process control, robotics and avionics. We will extend the methods for linear gaussian systems to a more generic problems using Monte Calro approach. The course will also cover multimodel approach and its use for the fault detection and isolation and introduction to adaptive control.
Study targets
No data.
Course outlines
1.Problem formulation, estimation methods
2.Bayesian approach to uncertainty description
3.Dynamic system model, probabilistic state definition
4.Identification of ARX model parameters
5.Tracking of time varuing parameters, forgetting, role of prior informaiton.
6.Numerically robust implementaiton for real time parameter tracking
7.Stochastic system, Kalman filter.
8.Kalman filtr for colour noise, extended Kalman filter, adaptive Kalman filter.
9.Stochastic dynamic programming, certainty equivalence principle.
10.Adaptive control, cautious and certainty equivalent strategies, dual control.
11.Probabilistic method for fault detection and isolation
12.Utilizaiton of multiple models
13.Nonlinear estimation, local approximation
14.Global aproximation, Monte Carlo Kalman filter
2.Bayesian approach to uncertainty description
3.Dynamic system model, probabilistic state definition
4.Identification of ARX model parameters
5.Tracking of time varuing parameters, forgetting, role of prior informaiton.
6.Numerically robust implementaiton for real time parameter tracking
7.Stochastic system, Kalman filter.
8.Kalman filtr for colour noise, extended Kalman filter, adaptive Kalman filter.
9.Stochastic dynamic programming, certainty equivalence principle.
10.Adaptive control, cautious and certainty equivalent strategies, dual control.
11.Probabilistic method for fault detection and isolation
12.Utilizaiton of multiple models
13.Nonlinear estimation, local approximation
14.Global aproximation, Monte Carlo Kalman filter
Exercises outlines
Laboratory covers work on individual assignments/projects.
.
.
Literature
No data.
Requirements
No data.
Responsible for the data validity:
Study Information System (KOS)