Autopilot Software Allows UAVs to Soar on Thermals – UAS VISION

A Navy scientist has re-engineered the software that allows long-endurance drones to powerlessly climb into the sky on bubbles of warm air. In a U.S. patent application published on May 2, Aaron Kahn, an engineer working on the Autonomous Locator of Thermals (ALOFT) project at the Naval Research Laboratory, reported that he has extensively tested the new software that detects and estimates the position of thermals, i.e., rising columns of warm air that birds use to stay aloft without flapping their wings. Unlike birds, soaring drones need the benefits of thermal detection and position estimation software as the warm air tends to drift relative to the ground due to winds. Prior systems relied on batch estimation processes that "require storing large arrays of data, which is not ideal for operation on small micro-controllers with limited memory resources." Kahn's new soaring software uses extended Kalman filtering, a kind of algorithm already used by the Navy for navigating submarines and cruise missiles. Now it can help orbit drones like the tiny CICADA glider or long-endurance solar-soaring UAVs that might also have photovoltaic or fuel cells feeding battery-powered propellers.