|
|
Estimating of motion models and its parameters to identify target trajectory
Benko, Matej ; Eliaš, Michal (oponent) ; Žák, Libor (vedoucí práce)
This text deals with removing noise from inaccurate multilateration measurements. It is used Bayesian estimation theory to find the posterior density of the real position (or, moreover, velocity) of an airplane. Together with true position, we estimate on Bayesian principle the geometry of a maneuver that an airplane obeys and so-called process noise, which describes how much an airplane's trajectory differs from the geometry. The estimation of the process noise is the essential part of the work. It is derived Bayesian approach together with the maximum likelihood approach. Then, improvements to these algorithms are introduced. They provide better results in particular cases, such as a maneuver change of the target or initial uncertainty of the maximum likelihood estimation. At the end of the text, the possibility of a combination of geometry and process noise estimation is described.
|
|
|
Comparison of filters in target tracking
Benko, Matej ; Bednář, Josef (oponent) ; Žák, Libor (vedoucí práce)
The topic of this bachelor thesis is Optimal Bayesian estimate usage in target tracking with bistatic measurement. The thesis is focused on particle filtering. It is shown particle filters are effective algorithms providing Optimal Bayesian estimate solution. There are tested and evaluated many types of fundamental algorithms, like SIR, Auxiliary or Regularized particle filters. It is compared to the accuracy of optimal estimates on various situations, depend on a different trajectory or acceleration of the target.
|
|
|
Estimating of motion models and its parameters to identify target trajectory
Benko, Matej ; Eliaš, Michal (oponent) ; Žák, Libor (vedoucí práce)
This text deals with removing noise from inaccurate multilateration measurements. It is used Bayesian estimation theory to find the posterior density of the real position (or, moreover, velocity) of an airplane. Together with true position, we estimate on Bayesian principle the geometry of a maneuver that an airplane obeys and so-called process noise, which describes how much an airplane's trajectory differs from the geometry. The estimation of the process noise is the essential part of the work. It is derived Bayesian approach together with the maximum likelihood approach. Then, improvements to these algorithms are introduced. They provide better results in particular cases, such as a maneuver change of the target or initial uncertainty of the maximum likelihood estimation. At the end of the text, the possibility of a combination of geometry and process noise estimation is described.
|
|
|
Comparison of filters in target tracking
Benko, Matej ; Bednář, Josef (oponent) ; Žák, Libor (vedoucí práce)
The topic of this bachelor thesis is Optimal Bayesian estimate usage in target tracking with bistatic measurement. The thesis is focused on particle filtering. It is shown particle filters are effective algorithms providing Optimal Bayesian estimate solution. There are tested and evaluated many types of fundamental algorithms, like SIR, Auxiliary or Regularized particle filters. It is compared to the accuracy of optimal estimates on various situations, depend on a different trajectory or acceleration of the target.
|
|
| |
host ::
RSS.