Estimating the attitude of spinning monocopters using low-cost sensors

What is it about?

Monocopters are tiny, single-wing drones inspired by spinning maple seeds—but their constant high-speed rotation makes it extremely hard for onboard sensors to figure out which way they’re tilted. Standard methods that rely on accelerometers fail because the spinning creates strong fake “gravity” signals that drown out the real one. In this work, we developed a clever workaround: instead of fighting the spin, we split the drone’s orientation into two parts—one describing the tilt of its spinning disk (like a helicopter’s rotor plane) and another describing how the body leans relative to that disk. By using only low-cost sensors (gyroscopes, magnetometers, and accelerometers) and an innovative algorithm that compares sensor readings over very short time intervals, we can filter out the spinning noise and recover the proper gravity direction. It lets us accurately estimate the drone’s whole attitude in real time—without needing expensive hardware or GPS. Our method was tested on a spinning platform and a real monocopter, demonstrating its reliability even during dynamic maneuvers. It paves the way for more autonomous, agile, and affordable bio-inspired flying robots.

Featured Image

Photo by Xiangkun ZHU on Unsplash

Photo by Xiangkun ZHU on Unsplash