In a remarkable leap forward for mechanical computing and soft robotics, scientists at the Flexible Structures Laboratory (fleXLab) at EPFL, together with researchers from AMOLF and Leiden University in the Netherlands, have unveiled a groundbreaking method to program mechanical metamaterials en masse, using nothing more than rotation. This innovative approach, called dynamic driving, leverages the physics of rotating systems to switch tiny elastic beams between two stable configurations, effectively encoding digital information directly into the material’s architecture.
The principle behind this inventive methodology resonates with a childhood favorite—the slap bracelet. This simple toy snaps from a straight band into a curled shape with a satisfying click. The snap arises from bistability, where the material has two energetically stable states. Borrowing from this concept, the research team has engineered bistable structures, which serve as mechanical bits (m-bits), capable of representing binary states 0 and 1 through their physical forms. Unlike electronic bits encoded via electrical charges, these m-bits harness mechanical deformation to store data, ushering in the possibility of entirely new forms of mechanical memory.
Historically, directly programming these bistable systems has been an arduous task. Each mechanical bit needed individual control, making the process tedious and impractical for large-scale applications. The dynamic driving innovation overcomes this intrinsic limitation by utilizing global rotation to simultaneously set the states of multiple bits. By finely tuning parameters such as spinning speed, direction, and acceleration, the team exploits inertial forces—centrifugal and Euler forces—that naturally manifest within rotating frames. These forces trigger snapshots of bistable beams, flipping them between stable positions in unison yet independently according to their distinct thresholds.
Central to this demonstration is a rotating platform loaded with five silicone beams, each roughly the size of a finger, affixed in a configuration that allows them to snap left or right. To illustrate the computational potential, each letter of the alphabet was represented as a unique five-bit binary string, derived from the standard ASCII encoding system. By calibrating the attachment points of each beam, researchers set different thresholds for beam flipping, dependent on the platform’s rotational parameters. As the platform spun, some beams crossed their mechanical snap points, flipping their state, while others remained static, cumulatively encoding a letter readable through the beams’ final orientations.
This intricate ballet of mechanical states was made possible by advances in motor technology. The team used a high-torque semiconductor motor, capable of extremely precise control over rotational dynamics, to finely orchestrate the motion. The synchronization of beam flipping to mechanical thresholds depended heavily on this precision. Eduardo Gutierrez-Prieto, co-first author of the study, notes that contemporary breakthroughs in motor performance were critical, enabling control dynamics that were previously unattainable for such soft, deformable systems.
Beyond the captivating demonstration of spelling the alphabet through rotation, the implications of dynamic driving extend far into emerging applications. Pedro Reis, head of fleXLab, emphasizes that this method isn’t limited to an experimental spinning platform but could revolutionize the way programmable mechanical metamaterials function in diverse environments. By embedding physical intelligence directly into materials, future robotic and mechanical systems could perform complex computations without reliance on conventional electronics, paving the way for smarter, more efficient devices.
One striking envisioned application lies in biomedicine, where centrifugal forces in spinning microfluidic devices could actuate bistable valves. This would enable precise and robust control of fluid flows in diagnostic systems, facilitating high-throughput medical tests without cumbersome external controls. Such mechanized valves, controlled purely through rotational dynamics, could greatly improve the scalability and portability of lab-on-a-chip technologies.
Soft robotics also stands to benefit significantly. Currently, soft robots often rely on embedded electronics and sensors, which add complexity and reduce resilience. With mechanical bits that respond to changes in pressure or rotation, robots can exhibit complex motion patterns autonomously, driven by simple pneumatic or hydraulic signals. This could foster the creation of more adaptable and environmentally robust robots capable of operating in extreme or remote conditions where electronics might fail.
From a theoretical perspective, the research enriches the understanding of how rotational inertial forces can be harnessed to control mechanical states in metamaterials. The technique elegantly illustrates the dual nature of the rotational frame—its ability to provide uniform, global energy input while still allowing for selective, local state changes in bistable systems due to variable snapping thresholds. It’s a fine example of applying classical mechanics principles in inventive engineering contexts.
Furthermore, the concept of mechanical memory embedded in physical structures resonates with the broader goals of embedding intelligence into materials themselves, a key ambition within the field of physical computing. Instead of carrying memory as abstract electronic signals, the memory lives literally in the shape and form of the material. This mechanical embedding of information is less susceptible to electronic interference, potentially offering more robust solutions for harsh or demanding applications, from deep-sea exploration to implantable medical devices.
The team’s work, published in Science Advances on May 6, 2026, represents an important merging of disciplines—mechanics, materials science, and information theory. By demonstrating that mechanical bits can be written dynamically and in parallel through rotation, they have unveiled a new paradigm for material intelligence and control. This breakthrough opens research avenues into remotely powered materials and devices that simultaneously compute, store memory, and actuate without onboard power or complex control circuits.
Looking forward, the dynamic driving method may inspire novel architectures for smart infrastructure that responds mechanically to environmental forces, underwater robotics that operate without electronics vulnerable to water damage, and microfluidic devices that self-regulate using simple rotational cues. Martin van Hecke of AMOLF highlights that the versatility of rotational control could enable new classes of devices spanning scales and disciplines, from tiny medical implants to large-scale autonomous systems.
In sum, this pioneering research transforms how scientists view materials—not just as passive substrates but as active computational entities. The innovative use of rotational inertial forces to program bistable mechanical bits represents a milestone toward embedded physical intelligence. As motor technologies continue to improve and the field of programmable metamaterials matures, the vision of smart, remotely operable mechanical systems is rapidly becoming a tangible reality.
Subject of Research:
Mechanical metamaterials, bistable structures, dynamic mechanical memory, and programmable materials.
Article Title:
Dynamic driving allows independent control of material bits for targeted memory.
News Publication Date:
6-May-2026
Web References:
DOI: 10.1126/sciadv.aec1606
Image Credits:
2026 fleXLab EPFL CC BY SA
Keywords:
Mechanical metamaterials, bistability, mechanical computing, dynamic driving, programmable materials, rotational control, physical memory, soft robotics, centrifugal force, microfluidics, material intelligence, mechanical bits
