Page 28 - august_072922.indd
P. 28
TAP TIMES -28- AUGUST 2022
Smart Socks (from 27)
To produce a smart textile, the researchers
use a digital knitting machine that weaves
together standard and functional yarn.
Multi-layer knit
The multilayer knit textile is composed of
two layers of conductive yarn knit sand-
wiched around a piezoresistive knit, which
changes its resistance when squeezed. The socks were developed and tested in the MIT Media Lab. (MIT)
Following a pattern, the machine stitches
this functional yarn throughout the textile “That thermoforming also allows us to cre-
in horizontal and vertical rows. Where the ate 3D forms, like a sock or shoe, that ac-
functional fibers intersect, they create a tually fit the precise size and shape of the
pressure sensor, Wicaksono explains. user,” he says.
But yarn is soft and pliable, so the layers Pressure sensor data
shift and rub against each other when the Once he perfected the fabrication process,
wearer moves. This generates noise and Wicaksono needed a system to accurately
causes variability that make the pressure process pressure sensor data. Since the
sensors much less accurate. fabric is knit as a grid, he crafted a wireless
circuit that scans through rows and columns
Knitting factory inspiration on the textile and measures the resistance
Wicaksono came up with a solution to this at each point.
problem while working in a knitting fac-
tory in Shenzhen, China, where he spent He designed this circuit to overcome ar-
a month learning to program and maintain tifacts caused by “ghosting” ambiguities,
digital knitting machines. which occur when the user exerts pressure
on two or more separate points simultane-
He watched workers making sneakers us- ously.
ing thermoplastic yarns that would start to
melt when heated above 70 degrees Cel- Deep learning techniques
sius, which slightly hardens the textile so Inspired by deep-learning techniques for
it can hold a precise shape. image classification, Wicaksono devised a
system that displays pressure sensor data
He decided to try incorporating melting fi- as a heat map.
bers and thermoforming into the smart tex-
tile fabrication process. Those images are fed to a machine-learn-
ing model, which is trained to detect the
“The thermoforming really solves the noise posture, pose, or motion of the user based
issue because it hardens the multilayer on the heat map image.
textile into one layer by essentially squeez-
ing and melting the whole fabric together, He and his colleagues are also exploring
which improves the accuracy. next page
