Grid Based Indoor Mapping Using Radar

Abstract

Burning buildings produce heavy smoke that hinders visibility indoors, reducing the effectiveness of conventional sensors like cameras and lidar, which are used by robots for mapping. Radar, on the other hand, is not affected by heavy smoke and it can keep reporting reliable measurements under adverse conditions. However, radars may suffer from sparse measurements and ghost objects which need extra measures to make sure that the radar is effectively utilized. In this work, the indoor mapping performance of an off-the-shelf commercial radar is evaluated, and a range based filtering method is proposed to reduce the effect of ghost detections.