MIT researchers have developed a soft sensor that recreates its own three-dimensional shape when it bends, folds and rotates. The technology could help remotely control robots, develop motion-tracking garments and measure patients’ progress during physical rehabilitation.
The system embeds flexible optical fibers in silicon and uses changes in transmitted light to track surface deformations. Software then converts these measurements into a digital reconstruction that tracks the movements of the material in near real time.
Unlike motion capture systems that rely on rigid sensors attached to clothing, the design uses soft components across the entire sensor surface. It can also tolerate damaged fibers without losing the ability to restore its overall shape.
Soft fibers track movements
The researchers created optical fibers from a flexible rubber core surrounded by a black rubber jacket. They deliberately roughened one side of each core to make the response to bending more obvious.
When a fiber bends, some of the light is scattered away from its path. The roughened surface makes this effect sensitive to the direction of bending, allowing researchers to estimate how each fiber bends by measuring the light emerging from its opposite end.
Individual fibers can detect curvatures along a line, but reconstructing an entire surface is more challenging. The MIT team addressed this problem by embedding multiple fibers in a stretchable silicone sheet.
Using computer simulations, the researchers were able to compare different fiber arrangements, including checkerboard and zigzag patterns. A particular zigzag configuration was found to be most effective at capturing complex changes in leaf shape.
Software reconstructs 3D shapes
Using this arrangement, the researchers created a physical sheet. LEDs send light through the fibers while sensors at opposite ends measure power. An external circuit board collects and amplifies the measured values.
An algorithm translates these measurements into a three-dimensional representation of the sheet metal. In demonstrations, researchers folded the material diagonally in opposite directions and its digital counterpart followed the changing geometry.
They also placed the plate over precisely shaped 3D printed molds to evaluate its accuracy. According to Qifan Yu, a mechanical engineering student at MIT, the reconstructed surface had an error of less than 0.4 centimeters.
Yu said comparable systems with rigid sensors typically produce errors of about 1 to 2 centimeters. The team also showed that the plate could restore its overall shape even after some fibers were cut or separated.
Wearable sensors could support therapy
The researchers envision garments that would allow users to control video game characters or operate robots remotely. The technology could also help physical therapists measure patients’ mobility during rehabilitation.
A therapist could wrap the material around a patient’s arm or leg to record movements and range of motion. Repeated measurements could help doctors compare mobility across sessions and assess recovery. Co-author Kaitlyn Becker, an assistant professor of mechanical engineering at MIT, said the technology could help therapists compare patient ratings over time.
The team now plans to make the fibers thinner. Their current fibers are about a millimeter in diameter, but improved manufacturing methods could reduce this thickness to tens of micrometers. Thinner fibers would allow researchers to embed more sensing elements into garments, potentially capturing finer details of body movement.
The study is published in the journal Advanced intelligent systems.
