Pattern tracking and visual servoing for indoor mobile environment mapping and autonomous navigation

In this paper, an image-based framework for avigation of a mobile robot in an indoor environment is presented. The only sensor used is an embedded monocular vision system. The environment is autonomously mapped during a learning stage in order to locate the robot on-line. A suitable control law is designed to drive the robot in an image database. To address this issue, a Virtual nonHolonomic Vehicle (VNHV) attached to the image plane is defined.