Authors:
Bach Van Pham
1
;
Simon Lacroix
1
;
Michel Devy
1
;
Marc Drieux
2
and
Thomas Voirin
3
Affiliations:
1
CNRS, LAAS;University of Toulouse, France
;
2
EADS-ASTRIUM, France
;
3
European Space Agency, Netherlands
Keyword(s):
Landmark Constellation, Pinpoint landing, Absolute navigation.
Related
Ontology
Subjects/Areas/Topics:
Computer Vision, Visualization and Computer Graphics
;
Motion, Tracking and Stereo Vision
;
Real-Time Vision
;
Visual Navigation
Abstract:
Precise landing position is required for future planetary exploration missions in order to avoid obstacles on the surface or to get close to scientifically interesting areas. Nevertheless, the current Entry, Descent and Landing (EDL) technologies are still far from this capability, as the landing point is predicted with a dispersion of several kilometres. Therefore, research has been conducted to solve this absolute localization problem (also called ``pinpoint landing''), which allows the spacecraft to localize itself within a known reference -- namely orbital imagery. We propose an approach (nicknamed ``Landstel'') which relies on Landmark Constellation matching that gives an alternative to the current solutions and also avoids the drawbacks of existing algorithms. The fusion of the inertial sensor relative motion estimation and the Landstel global position estimation yields a better global position estimation and a higher system's robustness. Position estimation results obtained bo
th with standalone Landstel and with the fusion of INS-Landstel via a simulator are shown and analysed.
(More)