Map building and path generation for Indoor mobile robot by using global and local views

Abstract

This paper proposes a novel method to generate an accurate and adaptive path for indoor mobile robots by using the global and local camera views. The three dimensional indoor map is necessary for the robots to navigate in the narrow rooms. Since the previous researches usually used the only camera or the sensors inside the robot, it was not easy to get a global map of the room for navigating. In the proposed system, the cameras with a fish-eye lens, attached along a line of the ceiling or at high-level positions on wall, provide the global views which can show the positions of both the moving robot and the fixed stuffs in the room. Also, the camera installed on the robot can provide the real-time local views, notifying the appearance of unexpected objects on its way and identifying the fixed stuffs like furniture in the room. Combing the global and local views can increase the probability of recognizing the objects around the robotic environment, and can provide more detail information of the three dimensional map for navigating.