Modern driver assistance and safety functions rely on the precise knowledge of the position and orientation of the ego vehicle. This is especially true for mapping approaches such as occupancy grid mapping. In this paper we describe a novel approach for selflocalization with six degrees of freedom. The measurement principle of scanning lidar sensors does not deliver proper point landmarks. Therefore scan-matchingbased approaches as well as point-landmark-based approaches show disadvantages in scenarios such as highways or tunnels. Our approach is capable of integrating the given information correctly and to improve the self-localization process. We use a feature-based approach using straight features, e.g. lines or planes, based on an extended Kalman filter.
Straight feature based self-localization for urban scenarios
Merkmalerkennung zur Selbstlokalisierung für innerstädtische Szenarien
2010
6 Seiten, 11 Bilder, 13 Quellen
(nicht paginiert)
Conference paper
Storage medium
English
Localization and Analysis of Critical Areas in Urban Scenarios
British Library Conference Proceedings | 2008
|