主催: 一般社団法人 日本機械学会
会議名: ロボティクス・メカトロニクス 講演会2023
開催日: 2023/06/28 - 2023/07/01
Conventional mobile robot navigation uses an accurate map of the environment, while humans can move with a rough map such as a hand-written map by comparing the map information and the shape of buildings. This research develops a method of matching a hand-written map with a map made by SLAM, aiming at hand-written map-based navigation. The method is composed of the following three steps: (1) wall detection by 3D voxel projection and probabilistic Hough transform, (2) building detection from detected wall lines, and (3) map matching by adjusting scaling and translation parameters. The method is tested with an actual SLAM map and a rough map.