Cooperative SLAM for Multi-Robot Systems using Visual Odometry and Range Measurements

Abstract

A swarm of robots can conduct exploration missions much faster and with increased robustness to isolated failures than a single robot. This dissertation introduces cooperative SLAM using visual odometry and ranges (CoVOR-SLAM) to successfully execute missions, even in the absence of satellite navigation support. Since CoVOR-SLAM only uses visual odometry and a few ranges, it requires much less computing power and communications than typical approaches using inter-agent loop closing.