Professor Antonio Forte of Oxford's Department of Engineering Science and Lead of RADLab commented, "We are excited to see that brain-less machines can spontaneously generate complex behaviours, decentralising functional tasks to the peripheries and freeing up resources for more intelligent tasks."
Soft robotics employ flexible materials ideal for negotiating uneven terrains and handling fragile objects. The principal goal is to embed decision-making and behaviour within the robot's structure to foster adaptability and responsiveness. Conventional electronic systems require dense sensor arrays and programming, making such physical intelligence hard to replicate.
To address this, Oxford researchers drew inspiration from biology, where tissues can serve multiple functions and synchronization appears naturally. Their main advancement was a compact modular block utilizing air pressure for mechanical tasks - as an electric circuit relies on current. This module can:
+ Actuate like a muscle by moving in response to air pressure
+ Sense pressure changes or touch
+ Switch airflow between ON/OFF, functioning as a valve or logic gate
Several identical units, each a few centimetres across, may be assembled like LEGO bricks to form various robots with unchanged hardware. Researchers demonstrated tabletop robots - about shoebox-sized - that could hop, shake, or crawl.
Specifically, each unit could simultaneously handle all three mechanical roles, producing rhythmic motion under constant pressure. When joined, these units naturally synchronized without external control or computers.
Such behaviours enabled creation of a shaker robot, which sorted beads by tilting a rotating platform, and a crawler robot able to sense table edges and halt to avoid falling. All coordinated actions arose mechanically, achieved without electronics.
Lead author Dr Mostafa Mousa of Oxford added, "This spontaneous coordination requires no predetermined instructions but arises purely from the way the units are coupled to each other and upon their interaction with the environment."
The synchronised motion only manifested when robots were both connected and grounded. Researchers used the Kuramoto model to describe how oscillator networks synchronize, finding complex motion emerged through physical coupling. The movement of each robotic limb affected its neighbours via shared frame and ground forces, creating feedback driven by friction, compression and rebound.
Dr Mousa explained, "Just as fireflies can begin flashing in unison after watching one another, the robot's air-powered limbs also fall into rhythm, but in this case through physical contact with the ground rather than visual cues. This emergent behaviour has previously been observed in nature, and this new study represents a major step forward towards programmable, self-intelligent robots."
Currently developed at tabletop scale, the underlying design principles are scalable. Oxford's team intends to investigate these dynamics for energy-efficient, untethered mobile robots, aimed at deployment in energy-constrained and extreme settings.
Professor Forte noted, "Encoding decision-making and behaviour directly into the robot's physical structure could lead to adaptive, responsive machines that don't need software to 'think.' It is a shift from 'robots with brains' to 'robots that are their own brains.' That makes them faster, more efficient, and potentially better at interacting with unpredictable environments."
Research Report:Multifunctional Fluidic Units for Emergent, Responsive Robotic Behaviors
Related Links
University of Oxford
All about the robots on Earth and beyond!
| Subscribe Free To Our Daily Newsletters |
| Subscribe Free To Our Daily Newsletters |
