The Simultaneous Localization And Mapping (SLAM) problem has been a matter of great importance and research in the area of intelligent robotics. The ability to map the environment and locate itself on the map simultaneously is an essential tool for mobile robots in an unknown environment. For localization, it is necessary to have maps. To map the surroundings, localization is needed. Very much like a chicken-and-egg problem. SLAM technology solves both the problem of localization as well of mapping together. Looking for answers to this challenge, different approaches have been developed, i.e. Visual SLAM (vSLAM), which is SLAM using cameras, in the case of this project, a RGB-D camera. In this Bachelor Project, the literature about robot navigation and the state of the art of SLAM approaches have been reviewed in deep. The system has been setup on the one hand, in simulation using Gazebo, and on the other hand, in a real a environment sys- tem; more precisely, using RSAIT’s Kbot in the first floor of the Faculty of Informatics (UPV/EHU). Experiments in both configurations revealed the potential of the tool for accurately mapping the environment avoiding odometry error, and allowed to learn the wide set of visualization tools available to ensure map correction and proper adjustment of some parameters. The obtained maps have been used later on to command navigation goals to the robot and to prove the usability of the learned maps.
Analysis of a RGB-D SLAM system using Real-Time Appearance-Based Mapping on Kbot
2022-06-17
128847-882457
Hochschulschrift
Elektronische Ressource
Englisch , Spanisch
DDC: | 629 |
Parsimonious Real Time Monocular SLAM
British Library Conference Proceedings | 2012
|