Mobile Robot Visual Control and Navigation Using Strategies of Moving Object Localization and Mapping

摘要

Mobile robots are navigated with a set of sensors including the visual sensors (for example PTZ or stereo cameras). The 2D or 3D images from these video sensors contain considerable information of the stationary and moving objects in the observation area of the robot. Therefore, many studies have been done to use this information for the purpose of control and navigation of a mobile robot. In general it is very difficult to solve this problem, but it is possible to facilitate the case, for example considering a room as the observation area of the mobile robot. This paper presents the visual model of a mobile robot system using a chosen strategy of mobile robot control and navigation. The robot is equipped with a sonar array for ensuring the obstacle avoidance and two cameras; one is a pan-tilt-zoom camera which is used for target detection and the other one is a bumblebee stereo camera on which the Simultaneous Localization And Mapping (SLAM) is based.