Researchers at the University of Cincinnati have developed a flapping-wing drone that can locate and hover around a light source without relying on artificial intelligence. The team, led by Assistant Professor Sameh Eisa from the College of Engineering and Applied Science, achieved this by using an extremum-seeking feedback system, a control method that enables real-time navigation through continuous performance-based adjustments.
The research, published in Physical Review E, demonstrates how the drone mimics the hovering behavior of insects such as moths and hummingbirds. The system adjusts flight parameters like wing-flap frequency to maintain stability and follow a target light source. Unlike conventional drones that depend on complex models or GPS, the extremum-seeking method allows the drone to stabilize and navigate autonomously using a simple, model-free feedback loop.
According to Eisa and doctoral student Ahmed Elgohary, the drone’s control mechanism reproduces insect-like hovering without computationally intensive algorithms. The design is inspired by biological flight, where insects make constant, small corrections to counteract environmental disturbances such as wind. The researchers found that their drone could replicate the distinctive hovering patterns of several species, including moths, bumblebees, dragonflies, hoverflies, craneflies, and hummingbirds.
The drone, built with four fabric-covered wire wings, was tested in a controlled environment. When operated manually, maintaining stable flight proved difficult, but the feedback system enabled the drone to hover autonomously, using slight, intentional wobbles to fine-tune its performance.
Eisa said the work contributes to understanding both bio-inspired robotics and the flight mechanics of small insects. He suggested that if hovering insects employ a similar feedback mechanism, the principle could have broader implications for biophysics and the design of future micro aerial vehicles.
Photo credit: Michael Miller
