Cooperative multi-robot monocular-SLAM using salient landmarks

In this paper we propose an algorithm to multi-robot cooperative Simultaneous Localization and Mapping (SLAM) with monocular camera mounted on each robot. The map consists of a sparse set of features, and the feature is prominent in the environment that robots explored, we call it salient landmark. The scene of exploration is typical office environment. Vertical lines and doorplates of environment are considered as visual landmarks to implement single-robot localization and multi-robot cooperation. An approach based on duplicate landmarks and rendezvous is also presented to improve the accuracy and reliability of SLAM. Our system was evaluated by an indoor two robot team, the results of this test showed that our approach is valid. ? 2009 IEEE.EI