MIT’s muscle-powered robot swims through a water maze under light control

MIT engineers say they have guided a thin, muscle-powered robot through a simple water maze by shining light on its fins. The demonstration, reported by MIT on September 29, shows how a small swimmer built with living cells can move and turn under external control. It took place in a petri dish, leaving proposed uses in natural waterways untested.

The robot’s body is a gel film roughly the length and width of a stick of gum. Its two halves act as fins, each coated with a layer of living skeletal muscle cells. The cells were engineered to contract in response to light, so illuminating one fin makes that side flap and pulls the swimmer through water. By choosing which fin to illuminate and changing the timing of the flashes, the researchers controlled its direction and speed.

What the maze test showed

For the reported maze test, the team submerged the robot in a large petri dish and manually moved a light source over it. MIT said the swimmer followed the light around a simple maze placed in the dish. This was a controlled demonstration of steering: the light source was moved by a researcher, and the published account does not describe an onboard navigation system or sensors carried by the robot.

The researchers reported straight-line swimming at up to four body lengths per minute. Their accessible preprint also reports turning at up to 1,200 degrees per minute under experimental conditions. Those figures describe the device in the team’s tests; they do not establish how it would perform in moving, dirty or otherwise less controlled water.

How the fins were made

The work builds on an earlier design from Raman’s group that used a thin layer of muscle cells on a grooved gel disk. In that device, light made the patterned tissue move in multiple directions, but MIT said the motion was about 100 micrometers. For the swimmer, the researchers focused on getting more useful movement by changing the support on which the cells grew.

The team tested the gel’s composition, stiffness and thickness, along with the shape of tiny grooves stamped into it. MIT reported that square-bottomed grooves aligned the cells better than curved ones. Better alignment helped the cells form muscle fibers that contracted together. A stiffer gelatin methacrylate gel also supported stronger contractions than the softer fibrin used in the group’s earlier design.

MIT said a gel film about half a millimeter thick provided support while remaining light enough for the muscle cells to stay attached as they contracted. The researchers also exercised the tissue with repeated light flashes before making the two-fin swimmer. The accessible preprint describes muscle films less than 15 micrometers thick on a patterned hydrogel support, placing the powered layer at a very different scale from the larger gel body.

In that preprint, the authors compare their design with previous biohybrid devices. They report ten times the force of earlier two-dimensional skeletal muscle actuators and 20 times the force per unit muscle volume of three-dimensional muscle tissue. They also report that an untethered actuator lasted more than 30 days, against less than 10 minutes for a previous two-dimensional design. These are the authors’ comparisons across designs and test conditions; the longevity result does not mean this swimmer operated independently in a natural waterway for a month.

Where the work fits

Living-muscle swimmers have a precedent. In a separate study published in Science in 2022, a Harvard-led team reported a biohybrid fish powered by cardiac muscle derived from human stem cells. That fish used muscle layers on opposite sides of its tail and an internal pacing mechanism to sustain swimming. MIT’s new design instead uses skeletal muscle on two fins and depends on externally directed light for steering.

MIT identifies Maheera Bawa as the new study’s first author and says the Office of Naval Research supported the work in part. MIT reported that the paper was appearing in Advanced Functional Materials on September 29; the accessible manuscript is a bioRxiv preprint posted in May. The journal version was not independently checked against a publisher record for this report.

Raman said the team’s next goal is to improve the body design for faster swimming. She also suggested environmental monitoring as a possible future use. For now, the demonstrated result is narrower: a muscle-powered robot that swam and turned under guided light in a lab dish. The reported work does not show it carrying monitoring equipment or navigating open water.