Autonomous Mobile Robot Navigation Considering the Pedestrian Flow Intersections

In recent years, many service robots move autonomously in an human co-existing environments. In these environments, there are multiple pedestrian flows and their intersections especially, in the indoor passage. At these intersections, the pedestrian’s movement tends not to be constant and for autonomous mobile robots, it is difficult to move safely with avoiding pedestrians. In our previous study, path planning considering the occlusion area in the indoor passage realized safe moving at the intersections. However, proposed system required intersection position in the environment beforehand. Therefore, the purpose of this study is development of a robot navigation scheme that realizes its safe movement for pedestrians at the intersection of their flow without special information beforehand. For realizing this objective, our proposed scheme estimates the intersection of the indoor passage without giving shape of the intersection to the robot beforehand. Using estimated results, our robot design the movement path, which is suitable for avoiding its risk at the intersection. The shape of the indoor passage can measure the confluence, but it is difficult to estimate the door that is closed state because the door has two different states: open and closed. In addition, the state of the door actually changes depending on movement of the pedestrian passing through the door, and the pedestrian on the passage side predict and avoid person that will appear from door changing the state of the door. The mobile robot can move the most suitable route by reflecting the walking characteristics of the pedestrian on the robot. In this study, we give type and position of door to the robot beforehand, and we propose the path planning using the walking characteristics of pedestrian for each door. In the simulation experiment and experiment using mobile robot, we confirmed that the proposed method is safer avoidance moving in front of the door.