Researchers working for Microsoft have created Capacitive, a smart fabric. Composed of sensors and an AI, it can recognize objects simply by placing it on them. What would be the applications of such an innovation?
Smart fabric to recognize many objects
You should know that smart textiles are not new in themselves. These include the fabric incorporating conductive fibers from the University of Massachusetts (United States) or the Jacquard connected jacket, resulting from a collaboration between Google and Levi’s. In a blog post published in October 2020, Microsoft researchers unveiled their innovation called Capacitive. It is a smart fabric prototype with a grid of twelve by twelve capacitive sensors.
How Smart Fabric works?
Capacitive can recognize a wide variety of objects, from food and dishes to plastic and paper objects. The system can recognize different fruits (kiwi, grapefruit, etc.). It can also make distinctions in liquids, whether it is hot or cold water, soda, or even milk. On the other hand, Capacitive is unable to recognize metallic objects or those with low permittivity. This is particularly the case with bank cards and even books.
Smart Fabric live
Improvements in anticipation
According to the researchers behind Capacitive, it is inexpensive to manufacture. Indeed, the sensors incorporate a conductive fabric that has been glued to a cotton substrate (non-conductive). The researchers cut the sensors into rows and removed the excess. Then, they connected the whole thing to an electronic circuit processing the signals using artificial intelligence. Also, the device uses random forests to recognize objects. This is a learning technique suitable for low power devices. According to the project directors, their tests show a recognition rate of 94.5% of the objects.
Final thoughts
As for the uses, the researchers seem to have an imagination. For example, Capacitive could suggest recipes incorporating the ingredients present on a table covered with the fabric or even warn when the user forgets his headphones on the same table. It could also monitor the water requirements of the plants. According to the researchers, the next step will be the integration of inductive sensors to recognize the metallic objects. Another improvement relates to the possibility of using the fabric as a simple touch surface controlling another device.
