Visual topological mapping and localisation using colour histograms

In this paper we present a system for appearance-based topological mapping and localisation using vision data. The algorithms are designed for robots which are equipped with FPGA~cameras. Such cameras do not provide the entire image to the robot but simple image features like colour histograms. Our mapping approach exploits the continuity of the visual appearance of consecutive images from the robots exploration traversal. For topologically mapping the environment colour histograms are clustered whereby each cluster represents a place. We use a Monte-Carlo localisation strategy combined with the topological map to localise the robot. For a robot equipped with a panoramic camera the proposed strategy works reasonably well, and is capable of overcoming the challenges of severe perceptual aliasing which occurs because of using simple image features and a sparse environment representation through the topological map.