Albatross UAV: Dynamic soaring

Albatrosses use dynamic soaring to cross thousands of miles of open ocean. This project aims to use dynamic soaring to overcome endurance limitations of small Unmanned Aerial Vehicles (UAVs). Previous research identified open loop control laws but for an efficient use with real aircraft, a closed loop control law is required. We intend to use intelligent control to identify such a closed loop control law using reinforcement learning. The second aspect of this project is skill transfer. Initially, for obvious safety and cost reasons, the controller training will be done using simulation. Afterwards, the learned control laws will be used with a real glider.