EPFL Engineers Create Sound-Powered Robots Using 3D-Printed Acoustic Cavities

At the Swiss Federal Institute of Technology Lausanne, engineers have figured out how to make machines move using nothing but sound waves and cleverly shaped hollow spaces. The breakthrough hinges on a principle that has been understood for centuries—the way air trapped inside a cavity will vibrate intensely when exposed to sound at just the right frequency. By designing these cavities with precision and manufacturing them through 3D printing, researchers at EPFL's MicroBioRobotic Systems Lab have created robots and flying devices that need no motors, no gears, no conventional electronics to propel themselves.

The work, published in Science Advances, takes an approach that inverts the usual relationship between sound and motion. Rather than using sound waves to push objects directly, the team engineered hollow structures—some round, some bell-shaped—that concentrate the oscillating air into a focused jet. The imbalance between this concentrated outgoing flow and the more diffuse air returning creates directional thrust. Selman Sakar, who leads the lab, describes it as transforming "a simple, cleverly designed mechanical piece into robotic matter." The cavities can be made from various materials: 3D-printing plastics, rubber-like polymers, even glass.

To demonstrate the concept, the researchers built miniature boats at the centimeter scale, each fitted with up to three cavities tuned to different audible frequencies. By adjusting the frequency emitted from a speaker, they could selectively activate individual cavities, steering the boats around obstacles and programming them for autonomous navigation. The same principle scaled down dramatically. Using 3D nanoprinting techniques, the team fabricated ultralight flying vehicles—microfliers—with microscopic cavities integrated directly into their polymer structures. One microflier, weighing just 150 micrograms, generated upward thrust through its cavities alone. Another incorporated the cavities alongside miniature blades that spun at speeds exceeding 13,000 revolutions per minute, producing aerodynamic lift.

What makes this approach particularly compelling is what it eliminates. Traditional robots require motors, bearings, magnetic components, and control circuits—all of which become increasingly difficult to miniaturize and integrate as devices shrink. By relying on acoustic resonance, the structures can be manufactured at scales that would be impractical with conventional mechanical systems. Junsun Hwang, a PhD student and first author of the study, notes that the concept remains compatible with even further miniaturization, opening pathways for designs that push the boundaries of what robotics and aeronautics can achieve.

The researchers envision a next phase in which multiple sound-responsive structures are integrated into a single flexible device, with each element tuned to respond to different frequencies. This would allow specific parts of a device to move, bend, or vibrate independently in response to acoustic signals. Sakar describes the potential: aerodynamic robotic devices that change shape on demand, their form and function dictated by sound. The implications extend beyond novelty. Devices that operate without onboard power sources or conventional actuators could function in environments where traditional robots would fail—inside the human body, in extreme temperatures, in spaces where electromagnetic interference is a concern. The absence of moving parts also means less wear, less maintenance, potentially longer operational lifespans. What began as a physics principle has become a manufacturing strategy, one that suggests the future of very small machines may be shaped by frequencies we can hear and others we cannot.