Researchers at MIT Media Lab have created a simple wireless system to sense potential food contamination, using RFID tags on billions of products. As food safety incidents have been one of the major headlines worldwide for causing illness and death over the past two decades, the new system called RFIQ aims to bring food-safety detection to the general public.
According to the researchers, a reader in RFIQ detects minute changes in signals emitting from RFID tags when the signals interact with food in containers. Although the study is initially focused on alcohol and baby formula, in the future, it may allow consumers to virtually sense the safety level of food before buying, they added. Further, it can be incorporated in smart fridges and back rooms of supermarket for continuous monitoring of RFID tags to automatically sense any spoilage of food.
The new technology works on the fact that particular changes in RFID-emitted signals correspond the level of food contamination within the product. Researchers modelled a machine-learning technique that learns correlation of the signals and the contamination level, thereby when given a new product, it can recognize purity of materials at different concentrations.
In experiments, the new system was 96% accurate in detecting melamine mixed in baby formula while the accuracy was 97% in alcohol diluted with methanol. Sensors usually developed for detecting chemicals and food spoilage are highly specialized systems, coated with chemicals to identify particular contamination. Co-author Fadel Adib said that main objective of the study to democratize food safety and quality and bring it to the hands of general public. With the new system, detection of food contamination has moved to computation side, allowing the cheap sensor to use for every product, he added.
The researchers explained that when RFID tags power up, they emit small electromagnetic waves that can be disrupted by the ions and molecules of the content of the product. This process is known as ‘weak coupling’ where changes in materials, alter the signal properties. They found out that feature distortions become more prominent with different materials and different contaminants. Materials can be classified on the basis of this information, showing varying characteristics between pure and impure materials.
Although RFIQ isn’t ready for commercial use, scientists at MIT are planning to produce significantly generalized technology that would work on wide range of products.