Going to the Extremum

Improving efficiency is one of the biggest challenges facing drone manufacturers today. The majority of autonomous aerial vehicles produced at present are quadcopters due to the incredible agility and speed they offer. But those capabilities come at the expense of energy efficiency — continuously spinning rotors drain a battery very quickly. This factor drastically limits the applications drones can be used for, so engineers are actively seeking solutions to the efficiency problem.

Researchers at the University of Cincinnati have proposed a two-pronged solution to this problem. First, rather than using rotors, the team took inspiration from nature. They observed that moths are both agile and efficient in flight, so they developed a drone that mimics their mode of flight. Furthermore, traditional flight control systems require a lot of battery-draining computations, so the team also developed an effective, yet minimal, control system for their drone.

The flight patterns of moths were replicated because they can remain stationary in turbulent air or follow a moving target by making extremely fast, fine-tuned adjustments — and they do all of this with very minimal “computational” power. The team’s observations led them to believe this is made possible by an extremum-seeking control mechanism, which allows for stable flight without artificial intelligence or complex modeling.

Extremum-seeking systems operate by continuously adjusting control inputs, such as wing flapping rate, based on immediate feedback about performance. This lets the system learn the optimal behavior in real time using a simple algorithm. The team’s flapper drone uses this principle to independently control roll, pitch, and yaw by flapping its four lightweight wings, each made of wire and fabric. To an observer, the wings appear as a blur, much like those of a hummingbird, but precise adjustments are always being made.

The drone’s control system measures its proximity to a target, such as a light, and constantly tweaks its motion to maintain the right position. A slight, intentional wobble in flight provides the necessary perturbations for the feedback loop — a feature seen in real insects. When activated, the drone can hover steadily, even replicating the subtle sway patterns of species like moths, bumblebees, dragonflies, and hummingbirds.

If extremum-seeking control does prove to be the mechanism insects use for hovering, it could reshape how scientists understand flight. And that could bring us one step closer to solving one of the biggest challenges in drone design — achieving stable, efficient flight without sacrificing agility.