MIT engineers build a thin swimming robot powered by light-controlled muscle cells

A thin layer of living muscle can do more than twitch: it can propel a small robot through water. MIT engineers demonstrated a flexible, two-finned swimmer whose genetically modified cells contract under light. The prototype swam and turned through a simple maze, showing that engineered muscle sheets can generate enough force to move a robot.

The study, published in Advanced Functional Materials, describes a gel framework roughly the length and width of a stick of gum. Each fin carries muscle tissue far thinner than a human hair. The experiment’s main goal was to establish that such thin tissue could power swimming, rather than to produce a finished underwater machine.

The muscle needed a better foundation

The project grew from an earlier artificial muscle disk in Ritu Raman’s MIT laboratory. Inspired by the human iris, that disk used patterned grooves to arrange muscle cells into a thin sheet. Light made the tissue stretch and compress in different directions, but its movements reached only about 100 microns—too small for the swimming task the researchers wanted to tackle.

One obstacle was the foundation beneath the cells. The earlier disk used fibrin, a very soft gel that could shrink under muscle contractions and limit useful movement. For the swimmer, the team tested the supporting gel’s chemistry, stiffness and surface geometry. Three design choices helped turn small contractions into stronger, coordinated motion:

1. Change the material. The researchers replaced fibrin with gelatin methacrylate, or GelMA, a gel used to grow engineered tissue. Adjusting its formulation let them compare different stiffness levels. Firmer surfaces produced better-aligned cells and stronger contractions.
2. Shape the channels. Square-bottomed grooves aligned muscle cells more effectively than the curved-bottomed channels the team tested. Cells growing in the same direction were more likely to fuse into fibers, helping the tissue contract together and produce force.
3. Balance support and weight. A gel film about half a millimeter thick supported a single layer of muscle cells while staying light enough to keep contracting cells from detaching. The foundation had to hold the tissue without becoming too burdensome for it.

The muscle also received a form of exercise before assembly. Repeated light pulses triggered contractions that strengthened the tissue. The finished swimmer had square-bottomed grooves and living muscle cells on both sides of its gel framework, with the cells developing into organized fibers.

Light supplies the steering

Each fin could be activated separately. Illuminating one caused that fin’s muscle to contract; illuminating both made both flap. Changing which fin received light, and how frequently, let the researchers adjust speed and direction. During the maze test, they moved the light source by hand over a water-filled petri dish. The demonstration therefore showed externally guided swimming, not independent route-finding.

Raman contrasts this thin design with biohybrid robots—machines combining living tissue and engineered structures—that use bulky, three-dimensional muscle grown from millions of cells. She says thinner designs could cost less to make and move more efficiently. Those are prospective benefits; this experiment focused on whether the thin muscle could generate sufficient swimming force.

The team’s next goal is to refine the body for faster swimming. Raman also sees possible uses in aquatic environmental monitoring and delicate settings where rigid machines could cause damage. Those remain proposed applications, beyond the simple dish-and-maze experiment. The Office of Naval Research partially funded the work.