Researchers at the University of Oxford have created a new class of soft robots that operate purely on air pressure — no electronic controllers, no motors, no computing hardware. These “fluidic robots” use modular mechanical units to produce complex, synchronized motion when several are linked and pressurized together.
The work addresses a key hurdle in soft robotics: embedding behavior and control directly into the robot’s physical body instead of relying on external electronics. Traditional soft robots still typically depend on sensors, microcontrollers and software to coordinate movement; by contrast, the Oxford team’s system embeds sensing, actuation and switching into a single repeating unit.
Each module functions like a mechanical building block. Under constant air pressure it can:
-
move or deform (acting like a muscle),
-
sense pressure changes or contact (acting like a sensor),
-
switch airflow on or off (acting like a logic gate).
Connected together like LEGO-style units, the modules generate rhythmic patterns of motion. When placed on a surface and pressurized, a linked set of these modules can hop, shake, or crawl — and their limbs naturally fall into sync via the interaction between ground reaction forces, internal structure and coupling between units.
In one example setup, a robot built with multiple modules automatically stopped walking when it detected the edge of a table — all without any electronic brain. The synchronization of movement was explained using the Kuramoto model, a mathematical description of how oscillators can synchronize when coupled.
From a design-engineering perspective this approach opens new possibilities:
-
By embedding decision-making into the structure itself you remove the need for many electronic components.
-
Using air pressure simplifies power and wiring since the mechanical structure handles coordination.
-
The modular design scales: the researchers point out that the principles apply across size scales, potentially enabling untethered robots in environments where power is limited or electronics fail.
Challenges remain. These prototypes still exist at tabletop size, tethered to air supplies. Long-term operation, ruggedization, environmental sealing and integration into larger systems are next milestones. But the underlying innovation is significant: a robot whose “brains” are in the body, not a chip.
For engineers working in fields such as soft robotics, environmental monitoring, or autonomous systems in extreme environments, this research presents a compelling design paradigm: build intelligence into the mechanics, not just the microelectronics.
