Aerial Robot Localization Using Ground Robot Tracking – Towards Cooperative SLAM

Heterogeneous ground-aerial robots cooperation becomes an increasing research area for autonomous tasks. Aerial robots are prone to a position drift over time; therefore, it is important to minimize it before achieving high level tasks. In this paper, a vision based position drift minimization method is proposed for cooperative navigation of quad-rotor and ground robot. The quad-rotor is equipped with a monocular camera and an IMU. The downward looking camera minimizes the translational drift of the quad-rotor by tracking the ground robot. After estimating the aerial robot's pose with respect to the ground robot, it can now localize itself into the map built by the ground robot using an EKF based SLAM algorithm. The proposed approach is implemented by simulating virtual robots in USARSim.