Wearable technology, which ranges from smartwatches that can measure all sorts of metrics for human movement and health to clothing that can monitor human vital signs without having to touch the wearer, is one area of development around the world. MIT researchers have developed clothing that detects a person’s movement through touch using special fibres. The smart fibres can tell if the person wearing the clothing is sitting, walking, or in a specific pose.
The clothing developed by MIT’s Computer Science and Artificial Intelligence Lab could be used for athletic training or rehabilitation. The innovative materials could also passively monitor health in assisted care facilities with the user’s permission, making it easier for semi-independent people to stay safe and alerting staff if the user falls.
Researchers at MIT have created a variety of wearable material prototypes, including socks, gloves, and full vests. To sense pressure from the wearer, tactile electronics use a combination of prevalent textile fibres and a small amount of custom functional fibres. According to one MIT researcher, developing a mass-production wearable with high accuracy data and a large number of sensors has traditionally been difficult. When multiple sensor arrays are manufactured, some of them will not work, and others will not work as well as others.
This motivated the team to develop a self-correcting mechanism that uses a self-supervised machine learning algorithm to detect and modify when those sensors deviate from their baseline. As the user switched from one pose to the next, the team’s socks were able to anticipate motion by looking at various sequences of tactile footprints and correlating them to different poses. The team’s vest can detect the wearer’s posture, movement, and the texture of the surface it’s touching.
Surprisingly, the research team claims smart clothing could be used in robotics to provide a sort of skin for robots to sense touch. The Toyota Research Institute helped finance some of MIT’s research. Antonio Torralba, Wojciech Matusik, and Tomás Palacios collaborated on the research paper with PhD students Yunzhu Li, Pratyusha Sharma, Beichen Li, postdoc Kui Wu, and research engineer Michael Foshey.