Ceiling vision based SLAM approach using sensor fusion of sonar sensor and monocular camera

Abstract

In practice, there could be more than two ceilings of different heights under an indoor environment and a robot under the desk could perceive the desk as a ceiling. In this case, a distance between ceilings and a robot seriously affects a performance of ceiling vision-based simultaneous localization and mapping (cv-SLAM). Therefore, in this paper, we propose the cv-SLAM using sensor fusion of mono-camera and sonar sensor. A sonar sensor can provide more exact distance between ceiling and a robot, and it helps to produce the improved performance of cv-SLAM.