Behaviour Based Multi-Robot Integrated Exploration
This paper presents an approach to the integrated exploration problem for a team of mobile robots. This technique is based on a combination of several basic behaviours that model a potential field. These behaviours are designed to quickly explore the environment jointly with a visual SLAM technique. As a novelty, this method considers returning to previously explored areas when the localization uncertainty is high. Consequently, the accuracy obtained in the construction of the maps is higher than with other classical exploration techniques. The known problem of local minima in potential field based techniques is also considered. In this sense, a strategy of detection and escape from local minima is used. Several simulations show the validity of the approach.