Abstract
In recent years, the development of autonomous
navigation technology has drastically improved. The increasing
prevalence of robotic systems has seen the emergence of a
range of navigation techniques, such as GPS-based navigation,
SLAM-based navigation, and monocular depth-based naviga-
tion. However, these techniques have their limitations, as they
rely on either external positioning systems or the need to build
a local map prior to navigation. This paper proposes a novel
approach to autonomous navigation: mapless navigation using a
pre-trained monocular depth network. This technique provides
an efficient and cost-effective means of navigation without the
need for a prior map of the environment
navigation technology has drastically improved. The increasing
prevalence of robotic systems has seen the emergence of a
range of navigation techniques, such as GPS-based navigation,
SLAM-based navigation, and monocular depth-based naviga-
tion. However, these techniques have their limitations, as they
rely on either external positioning systems or the need to build
a local map prior to navigation. This paper proposes a novel
approach to autonomous navigation: mapless navigation using a
pre-trained monocular depth network. This technique provides
an efficient and cost-effective means of navigation without the
need for a prior map of the environment
| Originalsprog | Engelsk |
|---|---|
| Titel | International Conference on Control, Decision and Information Technologiea |
| Antal sider | 6 |
| Forlag | IEEE |
| Publikationsdato | okt. 2023 |
| Sider | 1635 - 1640 |
| ISBN (Trykt) | 979-8-3503-1141-9 |
| ISBN (Elektronisk) | 979-8-3503-1140-2 |
| Status | Udgivet - okt. 2023 |