Dealing with visual features loss during a vision-based task for a mobile robot

Abstract

In this article, we address the problem of computing the image features when they become temporarily unavailable during a vision-based navigation task. The method consists in analytically integrating the relation between the visual features motion in the image and the camera motion. Then, we use this approach to design sensor-based control laws that are able to tolerate the complete loss of the visual data during a vision-based navigation task in an unknown environment. Simulation and experimentation results demonstrate the validity and the interest of our method.