Title :
High rate-localization for high-speed all-terrain robots
Author :
Roussillon, Cyril ; Lacroix, Simon
Author_Institution :
LAAS, Toulouse, France
Abstract :
Localization plays a central role in autonomous robot navigation, and a vast number of contributions can be found in the robotics literature. High speed all terrain robots raise new challenges for localization, that must run at higher rates, while being more accurate than for walking speed robots. The article presents a SLAM setup that satisfies these requirements, using vision and low-cost inertial sensors as its core. Other localization sensors can also be incorporated, such as GPS or odometry, yielding a system that estimate the robot pose and velocity parameters at 100 Hz, with translation errors on the pose of the order of 1m for 500m long trajectories.
Keywords :
Global Positioning System; SLAM (robots); mobile robots; path planning; sensors; velocity control; GPS; Global Positioning System; SLAM setup; autonomous robot navigation; high-speed all-terrain robot; inertial sensor; localization sensor; odometry; robot localization; robot pose parameter; robot velocity parameter; simultaneous localisation and mapping; translation error; walking speed robot; Cameras; Global Positioning System; Simultaneous localization and mapping; Uncertainty;
Conference_Titel :
Communications, Computing and Control Applications (CCCA), 2012 2nd International Conference on
Conference_Location :
Marseilles
Print_ISBN :
978-1-4673-4694-8
DOI :
10.1109/CCCA.2012.6417929