Dutch Researchers Build Brainless Soft Robot That Runs, Leaps, and Swims on Airflow Alone

A steady stream of air, with no instructions layered on top
The AMOLF robot's rhythm comes from airflow alone, without any electronic control system.
Mark

Why does a tube making a loud noise matter to robotics?

Mimi

Because Comoretto filmed it and saw a cycle sustaining itself. Pressure pushes a kink along, pressure drops, a new kink forms, and it repeats.

Mark

And that cycle becomes movement?

Mimi

That's the idea. The robot runs, goes over obstacles and swims, driven only by airflow, with no electronic control.

Luke

Careful with "brainless." The source says no central control system or AI. It doesn't say how the air supply itself is regulated.

Mark

Fair. What about the biological link?

Mimi

Comoretto noted the motion was asymmetric and periodic, which he says all living things share, from single cells to humans.

Luke

That's his observation of a pattern, not proof the robot moves like an animal. The gazelle and dog comparisons are the writer's imagery.

Mark

Is there any word on real-world use?

Luke

None in this piece. The suggestion of faster, simpler machines is a direction, not a demonstrated result.

Mimi

Right. What's confirmed is the mechanism and the three behaviors.

  • A robot with no processor and no electronic control is moving through the world anyway, which unsettles the assumption that movement needs a brain.
  • Conventional robots wait on central systems to react, and that delay leaves their motions stiff and mechanical.
  • Lead researcher Alberto Comoretto set out to trace a stray oscillating tube, filming it at high speed to understand the loop of pressure and kink.
  • The team is pushing the principle toward locomotion, testing whether airflow alone can carry a body across land, over barriers and through water.
  • The work is still early, and no timeline has been given for practical use, so its future remains open.

In an Amsterdam laboratory, a team at the AMOLF institute has shown that a machine can move with purpose without any mind at its center. Their soft robot runs, climbs over obstacles and swims, powered by nothing more than a steady flow of air. The discovery grew out of a small accident, a tube that refused to stay still, and it hints at an older truth: living things have long moved through rhythm rather than command, and perhaps machines can too.

A small team at AMOLF, the institute for molecular and atomic physics in Amsterdam, has built a soft robot that runs, climbs over obstacles and swims, and it has no brain at all. The only thing moving it is airflow. Most robots depend on a central electronic controller, whether or not it uses AI, and that controller takes time to respond, which is why their movements so often look stiff and mechanical. The AMOLF researchers wanted to show that a body could move gracefully without such a center.

The idea arrived by accident. Alberto Comoretto, the lead author, was working on another project in the lab and bent a tube to block airflow. Instead, the tube began to oscillate rapidly and loudly. Filming it with a high-speed camera, he found a self-sustaining loop: as a kink formed, rising pressure pushed it down the tube, the pressure dropped, and a new kink formed. The cycle repeated on its own.

What excited Comoretto was the rhythm itself. It was steady yet asymmetric, the same kind of pattern found in the movement of living things, from single cells to people. That resemblance pointed the team toward locomotion. The resulting machine demonstrates that a steady stream of air, with no instructions layered on top, can produce lifelike motion. Whether the design can scale beyond the lab remains the question to watch.

A robot that runs like a stotting gazelle and swims like a dog, and has no brain at all: that is the claim from a team of Dutch scientists at AMOLF, the institute for molecular and atomic physics in Amsterdam. Their soft robot can run, get over obstacles, and swim, and the only thing driving it is the flow of air.

The contrast the researchers are drawing is with the way most robots work. Typically a centralized electronic control system governs movement, whether or not it is powered by AI. Such a system needs time to react to changes in the environment, and the movements it produces are often awkward, in a word, robotic. The AMOLF team set out to show it does not have to be that way, building a machine with no electronic control and no central processor.

The idea began with an accident. Alberto Comoretto, a roboticist at AMOLF and lead author of the study, was in the lab working on another project. He needed to bend a tube to stop air from passing through it. Instead, the tube began oscillating at very high frequency and made a very loud noise.

Comoretto set up a high-speed camera to record what the tube was doing. The footage showed that the motion came from an interplay between the air pressure inside the tube and the physical state of the tube itself. When a kink formed, rising pressure pushed it along the length of the tube. That movement made the pressure drop, which allowed a new kink to form, and the cycle started again.

Comoretto told Ars that the team was "super excited" because they had seen self-sustaining, periodic, asymmetric motion. He had two reasons for the excitement. The first was that the flapping tube ran on the same kind of airflow physics that Peter Marshall, Doron Gazit, and Aireh Dranger used to build the dancing "Fly Guys" for the 1996 Olympic Games in Atlanta.

The second was the character of the movement. The asymmetry and periodicity in the tube's pattern are also present in the way all living things move, from single-celled organisms to humans. That resemblance is what pointed the researchers toward locomotion: a steady stream of air, with no instructions layered on top, was enough to produce a rhythm like that of living movement.

The source reports that the resulting robot runs, crosses obstacles, and swims, driven only by airflow. It offers no details here on practical deployment or timelines. The broader implication is for simpler machines that react faster because they do not wait on a central controller. How far the design can scale beyond the lab is the question to watch.

The team was super excited because they saw self-sustaining, periodic, asymmetric motion.
— Alberto Comoretto, AMOLF roboticist, to Ars Technica
Contact Us FAQ