3D Mapping for a Reliable Long-Term Navigation

Abstract

The use of maps allows mobile robots to navigate between known points in an environment. Using maps allows to calculate routes avoiding obstacles and not being stuck in dead ends. This paper shows how to integrate 3D perceptions on a map to obtain obstacle-free paths when obstacles are not at the level of 2D sensors, but elevated. Chairs and tables usually pose a problem when one can only see the legs with a 2D laser, although they present a high hurdle with a much larger area. This approach builds a static map starting from the construction plans of a building. A long-term map is started from the static map, and updated when adding and removing furniture, or when doors are opened or closed. A short-term map represents dynamic obstacles such as people. Obstacles are perceived by merging all available information, both 2D laser and RGB-D cameras, into a compact 3D probabilistic representation. This approach is appropriate for fast deployment and long-term operations in office or domestic environments, able to adapt to changes in the environment. This work is designed for domestic environments, and has been tested in the RoboCup@home competition, where robots must navigate in an environment that changes during the tests.