Optimal visual servoed guidance of outdoor autonomous robotic airships

This paper describes a visual servo control scheme for an outdoor autonomous airship in path following tasks. The visual based guidance takes full authority of the robotic airship and it is based on lines as image features, extracted from an on-board camera. An optimal controller is designed using the dynamic model of the system, which tackles the dynamics of the airship and the dynamics of the vision process. To prevent from losing image features during camera motion and to deal with noisy output measurements and modeling inaccuracies, a Kalman estimator was employed. The approach is validated through simulations using the underactuated airship of the AURORA project.