When soft robots learn to move without a brain
Alberto Comoretto defended his PhD thesis with the distinction cum laude at the Department of Mechanical Engineering on October 30th.
In nature, animals often move and react with a grace that seems instinctive. A jellyfish pulses through the ocean, a worm wriggles through the soil鈥攏one of them rely on complex brains to decide each motion. Instead, their soft bodies and the environment together create behavior. The PhD research of Alberto Comoretto shows that this same idea can be applied to machines. He explored how flexible materials and flowing air can give rise to autonomous, lifelike behaviors鈥攄irectly through physics, without the need for computers or control systems. The result is a new generation of soft robots that think with their bodies.
In most robots, memory lives in a computer chip. But what if memory could live in the material itself? By carefully shaping elastic shells and tubes, discovered that soft machines can 鈥渞emember past interactions. When touched or bent, they change shape in ways that persist鈥攁llowing them to switch behavior after an event. For example, one soft robot could turn itself in a new direction after bumping into a wall, purely because of how its materials were built. This physical memory makes machines programmable by design, not by software.
When valves start beating like hearts
Another surprising discovery came from a small fluidic valve that behaved in two ways at once: it could regulate pressure and produce rhythmic oscillations, similar to a heartbeat. The researcher modeled this using a mathematical framework that connects mechanical and fluidic forces. Understanding these dynamics allowed him to design an improved version that beats continuously鈥攊deal for powering devices such as soft artificial hearts that must never stop.
Tubes that walk themselves
Perhaps the most fascinating result came from a simple tube with air flowing through it. Instead of staying still, it began to oscillate and move鈥攁s if it had limbs. When several of these tubes interacted, they synchronized, producing coordinated movements that resembled walking or swimming. Even more remarkably, these soft 鈥渃reatures鈥� could avoid obstacles and switch gaits between land and water鈥攁ll without a brain or computer. Their behaviors emerged purely from the interplay between air, elasticity, and the environment.
Powered from within
To make these systems truly autonomous, Comoretto explored how they could draw energy from their own materials. By combining elastic membranes with a catalytic chemical reaction, he built a device that inflates and deflates rhythmically鈥攗sing energy released by the reaction itself. This marks a step toward machines that don鈥檛 need batteries or external pumps but instead live off their own internal chemistry.
A new kind of autonomy
Across all these experiments, one message stands out: complex, adaptive behavior can arise from simple physical rules. Nonlinear interactions between soft materials and fluids allow memory, motion, and synchronization to emerge naturally. This means the 鈥渂rains鈥� of future robots might not be computers at all鈥攂ut the physics of their bodies.
This work opens the door to artificial creatures that move, adapt, and survive through their design alone鈥攎achines that don鈥檛 just follow commands, but live through their materials.
Title of PhD thesis: . Supervisors: Dr. Bas Overvelde and Dr. Erik Steur.