Comparaison de Différents Filtres de Kalman et Modèles de Système pour la Navigation


  1. Beaudoin, Y.
  2. Gagnon, Eric
  3. Desbiens, A.
  4. Landry, R.
Corporate Authors
Defence R&D Canada - Valcartier, Valcartier QUE (CAN)
The work presented in this report was realized in the applied research project (ARP) 15eo04 and focuses on the navigation part. In navigation, an approach often used is to merge data from an inertial navigation system (INS) and a global positioning system (GPS) receiver using a Kalman filter. The purpose of this study is to analyze the computational load, accuracy of estimates and robustness to a bias on measurement according to the model and the Kalman filter selected. To do this, a train, moving along a two dimensional predetermined path, is used. Using a train instead of a satellite launcher is intended to simplify the analysis. Initially, the effect of grouping the states is studied. A short sampling period is used, thereby neglecting the effect of eliminating delays inherent in the grouping of states. Then, the states estimate is compared with the error states estimate. Finally, the extended Kalman filter is compared with the unscented Kalman filter.

Il y a un résumé en français ici.

Report Number
DRDC-VALCARTIER-TM-2011-184 — Technical Memorandum
Date of publication
01 Apr 2012
Number of Pages
Electronic Document(PDF)

Permanent link

Document 1 of 1

Date modified: