Optical-Flow Based Strategies for Landing VTOL UAVs in Cluttered Environments

Abstract

This paper considers the question of landing an Unmanned Aerial Vehicles (UAV) using a single monocular camera as the primary exteroceptive sensing modality. The proposed control law is based on tracking a single point feature, representing the desired landing point on a ground plane, along with optical flow computed over the full image. The bearing of the desired landing point is used as a driving term to force convergence, while the optical flow is used to provide a damping force that guarantees both obstacle avoidance as well as damping the convergence of the vehicle to the ground plane ensuring a soft touchdown. A detailed analysis of the system closed-loop dynamics is undertaken and the response of the system is verified in simulation.