Project Albatross: Bio-Inspired Reactive Control Strategies for Atmospheric Energy Harvesting in Fixed-Wing UAVs
The Project
Project: ALBATROSS is a bio-inspired autopilot that teaches a small RC glider to “ride” the wind like an albatross and stretch every watt of battery power. What it does: Instead of holding a fixed speed or altitude, the controller looks for tiny energy opportunities in low-altitude wind shear and gusts. It times short pitch changes and small throttle taps so the glider climbs while crossing faster air and descends through slower air, repeating a gentle figure-eight pattern. I measure success with energy metrics such as watt-hours per kilometer and net potential energy gain, both in simulation and in field tests on a soccer field. What inspired it: Albatrosses travel huge distances with almost no flapping. Nature already solved the endurance problem with dynamic soaring. I studied that flight pattern, treated evolution as a patient scientist, and translated those ideas into algorithms I could test with a dual-motor glider, a Pixhawk, and onboard sensors. What was difficult: Three things stood out. Wind modeling: creating realistic, fast-changing gust and shear models that still run in real time. Sensing and timing: extracting clean cues from noisy IMU, GPS, and airspeed data, then triggering millisecond-scale maneuvers at the right moment. Closed-loop tuning outdoors: validating energy gains safely on a small field with changing winds and strict battery limits. Why it matters: If small aircraft can harvest ambient energy intelligently, they can fly farther with less battery weight and lower cost. The same ideas can help environmental monitoring drones, hobby aircraft, and future autonomous gliders.
