Wing-strain-based flight control of flapping-wing drones through reinforcement learning

Although drone technology has advanced rapidly, replicating the dynamic control and wind-sensing abilities of biological flight remains challenging. Insects use wing-mounted mechanoreceptors (campaniform sensilla) to detect complex aeroelastic loads essential for agile flight. Using robotic experiments that mimic these biological systems, we show that wing strain encodes critical information about a drone’s attitude, as well as wind direction and speed. We introduce a wing-strain-based flight controller that infers attitude and airflow from aerodynamic forces on flapping wings, eliminating the need for accelerometers and gyroscopes. Across five experiments—sensor validation, single-DOF control under changing winds, two-DOF attitude control via gravity-based adjustment, position control in windy conditions, and precise flight path control in calm air—we demonstrate that a flapping drone can be controlled using only wing strain sensors coupled with a reinforcement-learning-based flight controller. The system adapts to environmental changes, supporting applications from gust resistance to wind-assisted autonomous flight.

Taewi Kim, Insic Hong, Sunghoon Im, Seungeun Rho, Minho Kim, Yeonwook Roh, Changhwan Kim, Jieun Park, Daseul Lim, Doohoe Lee, Seunggon Lee, Jingoo Lee, Inryeol Back, Junggwang Cho, Myung Rae Hong, Sanghun Kang, Joonho Lee, Sungchul Seo, Uikyum Kim, Young-Man Choi, Je-sung Koh, Seungyong Han, Daeshik Kang