Researchers from the Massachusetts Institute of Technology (MIT) have developed smart textiles that snugly conform to the body and can sense the wearer’s posture and motions using a novel fabrication process. The researchers were able to greatly improve the precision of pressure sensors woven into multi-layered knit textiles tag 3DKnITS, by incorporating a special type of plastic yarn and slightly melting it with heat, a process known as thermoforming.
“With digital knitting, you have this freedom to design your own patterns and also integrate sensors within the structure itself, so it becomes seamless and comfortable, and you can develop it based on the shape of your body,” Irmandy Wicaksono, research assistant, MIT Media Lab and lead author.
They used this method to develop a “smart” shoe and mat and then developed a hardware and software system to measure and interpret data from pressure sensors in real-time. The machine-learning system accurately predicted motions and yoga poses performed by a person standing on the smart textile mat.
Their fabrication process, which makes use of digital knitting technology, allows for rapid prototyping and is easily scaled up for large-scale production. The technique has a wide range of potential applications, particularly in health care and rehabilitation. It could be used to make smart shoes that track a person’s gait as they learn to walk again after an injury, or socks that monitor pressure on a diabetic patient’s foot to prevent ulcers.
The researchers use a digital knitting machine to weave together layers of fabric with rows of standard and functional yarn, but because the yarn is soft and pliable, the layers change and squeeze as opposed to each other when the wearer moves. This causes noise and variability, making the pressure sensors much less accurate.
Moreover, the researchers decided to experiment with melting fibres and thermoforming in the smart textile fabrication process. Thermoforming effectively eliminates noise because it hardens the multilayer textile into a single layer by squeezing and melting the entire fabric together, improving accuracy.
This thermoforming also allows for the creation of 3D forms, such as a sock or shoe, that are precisely tailored to the user’s size and shape. Following the fabrication process, researchers required a system to process pressure sensor data accurately.
Researchers created a system that visualises pressure sensor data as a heat map, drawing inspiration from deep-learning methods for picture classification. A machine-learning model is trained to recognise the posture, position, or motion of the user based on the heat map image using the photographs that were provided as input.
After the model had been trained, it could identify seven yoga positions with 98.7 per cent accuracy and characterise the user’s behaviour on the smart mat (walking, jogging, doing push-ups, etc.) with 99.6% accuracy.
A form-fitting smart textile shoe with 96 pressure sensing spots dispersed across the full 3D textile was also made using a circular knitting machine. The shoe was designed to gauge the amount of force applied to various foot regions during a soccer ball kick.
Since precision is crucial in prosthetic applications, 3DKnITS’ great accuracy may be advantageous. Additionally, researchers are looking at new original uses.
After demonstrating the success of their fabrication technique, the researchers intend to refine the circuit and machine learning model. Currently, before the model can classify actions, it must be calibrated for everyone, which is a time-consuming process.
Eliminating the calibration step would make 3DKnITS more user-friendly. The researchers also intend to conduct tests on smart shoes outside of the lab to see how environmental conditions such as temperature and humidity affect sensor accuracy.