In English

Simultaneous localization and mapping for vehicles using ORB-SLAM2

Marcus Andersson ; Martin Baerveldt
Göteborg : Chalmers tekniska högskola, 2018. Master's thesis - Department of Mechanics and Maritime Sciences; 2018:85, 2018.
[Examensarbete på avancerad nivå]

This thesis details work developing a containerized version of ORB-SLAM2 implemented in the open source software framework OpenDLV as well as testing it both on established datasets as well as on gathered data in a real traffic scenario in the Gothenburg area. Parallel to the real traffic scenarios this thesis also evaluates the feasibility for ORB-SLAM2 to be run on a formula student race car, due to the authors involvement in the Chalmers formula student driverless project which will compete with an autonomous race car in the formula student germany competition. The thesis also evaluates ORB-SLAM2 as a SLAM solution on a regular car and the strengths and weaknesses of the system in that area.

Nyckelord: Localization, mapping, SLAM, ORB-SLAM visual odometry, autonomous vehicles, stereovision, OpenDLV, 3D reconstruction



Publikationen registrerades 2018-11-12. Den ändrades senast 2018-11-12

CPL ID: 256291

Detta är en tjänst från Chalmers bibliotek