Pressure-sensing mats and shoes could enhance healthcare and video games

MIT researchers created a smart textile linked up with a machine learning system.
MIT 3D knits
MIT's smart textiles can sense how you move. Irmandy Wicaksono / MIT

The clothes you wear are not smart—the cotton and synthetic thread blends lack any kind of sensors in them. But scientists at the Massachusetts Institute of Technology have a different kind of material in mind.

MIT researchers used a uniquely designed plastic yarn to create a knit textile, called 3DKnITS, that they interspersed with pressure sensors. In their experiments, they used this material to construct shoes and mats. They also paired it with a hardware and software system (including a machine learning component) that measured and interpreted the incoming pressure sensor data, and used that to predict a person’s movements. Their design will be described at the IEEE Engineering in Medicine and Biology Society Conference.

Smart textiles that can sense how users are moving could be useful in healthcare, for example, for monitoring gait or movement after an injury. Athletes could donn them to receive feedback on their movements, or they could be used to create a better video game interface. 

During their tests, researchers connected the textile to a Minecraft video game, using it as a controller to move through the virtual world. It could detect whether the user wanted to shuffle left or right depending on what foot they were standing on as well as if they wanted to jump, walk, or run.

[Related: This wearable fabric microphone can listen to the world—and your body]

They also used the mats as part of a yoga practice, so they can see how well it detected postures like eagle, tree, or warrior, based on the pressure distribution it was sensing on the textile surface.  

For the fabric, a pressure sensor is active at each point where two threads intersect. A wireless circuit scans and measures the force applied at each sensor. The pressure input is displayed as a heat map on a linked up computer screen, and the image of that is fed to a deep-learning system, which has been trained to use the heat map to predict posture, pose, or motion. After training, the researchers claim that the system is able to classify physical activities like walking, running, and push-ups with 99.6 percent accuracy. It’s also able to nail down seven yoga poses with 98.7 percent accuracy. 

The team claims that they can also form these textiles into socks, sleeves, and more. They could also make custom fits by 3D scanning, or 3D printing, parts of the human body, and then steaming or heating the textile to form it into the desired shape. In another creative endeavor, the team made the textile into a magic carpet of sorts, one that curates music that shifts in soundscapes in accordance to a dancer’s steps. 

Watch the textile in testing below: