Landmark Design Using Projective Invariant for Mobile Robot Localization

Abstract

For the fast and accurate self-localization of mobile robots in navigation, artificial landmarks can be used very efficiently in the complex workspace. In this paper, in order to provide feedback and verification mechanisms in navigation technique, we design two types of artificial landmarks with symmetric rectangles and seven-part numbers, which show cross- ratio invariant under projective transformation. The fast landmark detection-recognition algorithm and self-localization are proposed and their feasibility and robustness are demonstrated by practical experiments in cluttered indoor environments and simulation experiment. Experimental results show that proposed landmark patterns are enough to be used in cluttered environment and landmark detection-recognition in cluttered scene is real-time robustly under various viewing angles, self-localization accuracy is high enough in the presence of additive random noise.