抄録
This paper proposes a novel method to easily and effectively teach a mobile robot its path and to navigate it along the instructed path. First, an operator takes pictures at some significant points such as a start point, turning left/right points and a goal point at the teaching stage. At the same time, locations of these significant points are also obtained with a GPS receiver. Then, the operator teaches the robot its path by some simple instructions based on the pictures referred as reference images. Finally, at the autonomous navigation stage, the robot travels along the path with two navigation modes. One of them is landmark-based navigation, that is, the robot recognizes its headings and the significant points by detecting the landmarks indicated in the reference images when the robot is traveling near the significant points. Another mode is GPS-based navigation. The robot is navigated to near the significant points by GPS information when the robot is far from the significant points. Experiment conducted along a 620 meters long path indicates the effectiveness of the proposed method.
