Robotic System for Mapping 3D in-wall Information for Craftsmen

This paper presents a robotic system for autonomously scanning wall surfaces by means of inductive, capacitive and AC measurements in order to gather information about flush-mounted power lines, water pipes and cavities. From these data the system generates a 3D map with in-wall information. Algorithms for surface scanning are described enabling robust methods for data acquisition and fusion of the scanning and localization data. The paper describes and evaluates two methods for reconstructing surface and in-wall information: The first one uses occupancy grid mapping along with the elaborated sensing model of the wall scanner, the second is a combination of scanning and mapping. The created 3D map is made available to a second system that projects the map onto the wall surface, removing distortions induced by the lack of a perpendicular projection. That system provides craftsmen the additional information to prevent hitting wires or water pipes when performing drilling tasks. The utilized robots make use of the robot software framework ROS. Keywords: Autonomous wall scanning, in-wall information, flush-mounted piping and electric installations, sensor fusion and localization to a 3D-map, projection of in-wall information, SLAM, occupancy grid mapping, ROS