Autonomous Maritime Landings for Low-cost VTOL Aerial Vehicles

Autonomous landing of quadrotor UAV on a maritime vessel is a challenging task, as low cost sensors, unknown movements of the landing surface, and external disturbances make it difficult to generate a relative pose estimate with sufficient accuracy for landing. In this work, we propose an architecture that avoids sensor limitations while allowing for accurate relative pose estimation, even in the presence of wind disturbances. The final landing sequence is performed entirely in the body-fixed inertial frame so that noisy measurements from the GPS and magnetometer sensors do not degrade the relative estimation accuracy. Simulation results of the entire system architecture are presented, as well as experimental results of visual landing pad tracking for representative motions, which demonstrate the validity of the approach.