Visual feedback for autonomous navigation

Abstract

The author addresses the problem of visual navigation, and through it the more general problem of sensor-mediated movement. A new approach to the navigation problem is proposed, in which visual processes are closely and actively integrated with the control of the robotic system. In the case of obstacle avoidance, the concepts of free space doors and corridors are introduced and it is shown that this dynamic information is easier to extract from the sequence of incoming images than the depth map traditionally used. Though poorer in content, the information is still sufficient for the purpose of closed-loop planning and control of trajectories to the goal.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>