Descripción
|
|
---|---|
?In this paper, an accurate, efficient, and simple vision-based pose estimation strategy for UAV navigation in GNSS-denied environments is presented. Using visual information and previous knowledge of 3D geometries present in the environment, the pose can be estimated accurately and used for autonomous navigation. The indoor mission in the IMAV 2016 competition has been chosen for developing and evaluating this approach. Three Perspective-n-Point (PnP) algorithms have been tested and benchmarked with the purpose of selecting the most suitable for navigating in this scenario. All of them have been tested in a realistic Gazebo-based simulation using our novel UAV software, Aerostack, which allows for a fully autonomous solution. A complete flight in a GNSS-denied environment has been successfully simulated, indicating that real flights are feasible with this approach | |
Internacional
|
Si |
Nombre congreso
|
International Micro Air Vechicle Competition and Conference 2016 |
Tipo de participación
|
960 |
Lugar del congreso
|
Beijijng, China |
Revisores
|
Si |
ISBN o ISSN
|
- |
DOI
|
|
Fecha inicio congreso
|
21/10/2016 |
Fecha fin congreso
|
24/10/2016 |
Desde la página
|
22 |
Hasta la página
|
27 |
Título de las actas
|
- |