屋内自律移動ロボットのための環境の特徴的方位の検出手法

抄録

The current orientation estimation is the principal task for mobile robots. We have been developing the orientation estimation method for indoor mobile robots, using the properties of the ordinary building structure, which we can assume that most of walls and furnitures are arranged in parallel or perpendicular to the dominant direction. Based on this assumption, robots can detect the characteristic geometry of environment efficiently. In this paper, we propose two kinds of methods to find the dominant direction of the building from the surrounding range data based on these characteristics, to estimate the orientation of the robot by the scanned range data. And we present the results of experiments to verify the effectiveness of the proposed methods.