Turn Decision-Making for Improved Autonomous Thermalling of Unmanned Aerial Gliders

Unmanned Aerial Vehicles (UAVs) are increasingly being proposed and used for a wide range of civilian applications. A subclass of fixed-wing UAVs, known as Unmanned Aerial Gliders (UAGs), are particularly well-suited to applications involving long endurance and range. An ongoing challenge to achieve full autonomy for these aircraft, is to improve the detection and exploitation of rising air masses (thermals). This paper proposes a turn decision-making algorithm for improved thermal exploitation with UAGs. The algorithm monitors netto-variometer data composed of vehicle states to identify the presence of thermal updraft. Following thermal detection, a turn decision into the thermal is made based on the induced roll effects on the UAG due to updraft wind gradients. Once inside the rising air mass, the thermal center location is estimated using the centroid method. The effects of turn decisions on thermalling performance are initially examined in a batch simulation involving various thermal properties and wind conditions, implemented in MATLAB/Simulink. The algorithm is subsequently integrated into the PX4 flight stack for flight testing and results demonstrate the performance of the algorithm and improved thermalling when making intelligent turn decisions to capture a thermal.