Autonomous Locator of Thermals (ALOFT) Autonomous Soaring Algorithm.

摘要

The Autonomous Locator of Thermals (ALOFT) algorithm goal was to develop an algorithm which could exploit naturally occurring convective thermal updrafts for extending the endurance of an unmanned aerial vehicle (UAV). A commercial off-the-shelf SBXC sailplane was outfitted as a UAV and used for more than 100 test flights of the ALOFT algorithm, with a nominal endurance of 3 minutes after a winch-launch to 100 meters. A notable success was unofficially breaking the cross-country soaring goal-and-return world record by flying 60.4 total miles over approximately 4.55 hours. Best endurance demonstrated by the algorithm was 5.3 hours and best range demonstrated by the algorithm was 70.47 miles. This report documents the final ALOFT software algorithms developed, to capture the most important implementation details and mathematical formulas. Specifically, the soaring algorithm is divided into four major functions: reading state data from the autopilot, identifying the position and other characteristics of a nearby thermal, making soaring behavioral decisions, and sending soaring commands to the autopilot.