Detecting Food Contamination Using the RFID Tags Already on the Containers

Food contamination is a serious issue, particularly in countries with less strict regulation than the United States. In 2008, 50,000 babies in China had to be hospitalized after being fed melamine-contaminated formula. In Indonesia this year, more than 100 people were killed after drinking alcohol that had been diluted with methanol. Government regulations and penalties should be strong enough to prevent that from happening, but implementing that is a lot easier said than done. In the mean time, MIT researchers have come up with a way to detect contamination using RFID tags.

RFID tags are very inexpensive and are already placed on the packaging of many food and drink items, so this technique could be easily deployed without incurring prohibitive costs. All consumers would need to do is use a specialized RFID reader to determine if a product is contaminated. This doesn’t work by looking up the product in a database to check if it’s known to be contaminated — that would still carry a lot of risk and would require significant resources.

Instead, this works by using the RFID tag itself as an actual sensor that can detect many types of contamination. That’s possible because each RFID tag acts like a rudimentary radio frequency spectroscope. Radio frequency spectroscopy works by projecting EM (electromagnetic) waves into a material, and then monitoring how those waves are affected in order to determine the composition of the material.

MIT’s technique works the same way. The reader emits a 950MHz signal to activate the tag, and then emits secondary signals ranging from 400MHz to 800MHz to identify the material inside a container. A machine learning system is trained on the reflected signals from known samples, and can then identify contaminates. During tests, they could detect melamine-tainted baby formula with 96% accuracy and methanol-diluted alcohol with 97% accuracy.