Image credit: Dartmouth College/XDiscoveryLab
The Internet of Things allows people to monitor all kinds of health-related factors, from steps taken to respiratory distress. Mobile apps enable users to count calories, track food consumption, and suggest recipes. In the future, a tablecloth might be able to do the same.
Researchers at Microsoft and Dartmouth College’s XDiscoveryLab are developing a capacitive-sensing technology that enables fabric with embedded IoT to detect objects placed upon it. The Capacitivo textile can identify 20 non-metallic household objects, such as food items and dishware. Capacitivo also is able to identify whether a glass is full or empty, what is inside, and sends that information to an app.
Interactive Fabric with Object Recognition
The Capacitivo textile uses machine learning and a thin grid of diamond-shaped conductive electrodes that is heat-bonded to a standard textile fabric. A layer of conductive fabric sits on the underside of the piece as a ground plane. When an object is placed on Capacitivo, the fabric senses the shift in capacitance due to the object’s permittivity. Each object has a different permittivity, so Capacitivo measures the changes in capacitance to recognize the object. Machine learning helps determine the size, shape, and material of a specific item.
Initially the researchers tested 20 items, including an avocado, kiwi, candle, earbud case, and credit card. They placed each item on the fabric ten times in random order to “train” the fabric. The Capacitivo textile was later able to recognize the item in real time with an accuracy of 94.5 percent.
In a test of liquids, Capacitivo was able to distinguish between milk, apple juice, soda, beer, hot and cold water with 90.7 percent accuracy, though it is less accurate distinguishing between soda and beer. It has other limitations. Capacitivo does not work well with square-edged objects, such as books, nor does it detect metal objects.
Find my Phone: Metal-Sensing Textiles
The Dartmouth team previously developed a similar technology, Tessutivo, an inductive-sensing fabric that can recognize metallic objects, such as a phone, metal water bottle, spoon, soda can or coins. In tests, Tessutivo was able to identify 27 objects in real-time with 93.3 percent accuracy.
The fabric consists of four layers. The middle two layers of fabric contain stitched rectangular-shaped coils. Those are sandwiched between two insulating outer layers. When a metal object rests on the coil, it creates an electromagnetic field coupling between the sensor coil and the object. Tessutivo identifies each object based on the couplings and machine learning.
Tessutivo can trigger alerts to remind users to take action or to prevent lost objects. For example, a water bottle placed on the fabric will automatically start a timer that will notify the user when to take a drink. Tessutivo can also sound an alarm indicating a phone remains on the couch after a person stands up, so they don’t walk away without it. It also can tie into a smart home device that could notify families when dinner is on the table.
Researchers predict that Capacitivo eventually will help home cooks by suggesting recipes based on identified items or prompting cooks to add ingredients during food preparation. The technology could be tied to other smart systems, allowing it to upload data to food or diet tracking apps.
In the future, Capacitivo could alert users if they leave a credit card or ear buds on the table or notify them when to water a plant. If these fabrics take hold, it could lead to better health, less food waste, and fewer misplaced items.