An autonomous mobile robot in a human's living space should be able to realize not only collision-free motion, but also human-centered motion, i.e., motion giving priority to a moving human according to the situation. In this study, we propose a real-time obstacle avoidance method for an autonomous mobile robot that considers the robot's dynamic constraints, personal space, and the human's directional area using grid-based X-Y-T space path planning. The proposed method generates collision-free motion in which the robot can give way to humans. The relative position, velocity and avoidance motion with respect to the robot varies from person to person. To show the effectiveness of the proposed method to human-like motion, we verify the robot's motion under several assumed scenarios by changing the initial state of both the robot and the human. Moreover, we verify the robot's motion with respect to a simulated human based on the human-like behavior approach. Through these simulations, we confirm that the proposed method is able to generate safe human-centered motion under several assumed scenarios. Additionally, the effectiveness of the proposed method in practice is confirmed by experiments in which the human's position and velocity are estimated using a laser range finder.
ASJC Scopus subject areas
- コンピュータ サイエンスの応用