Ohio State AI e-nose sends smells to a wearable Scentronic device

Researchers at Ohio State University have announced a system in which an AI electronic nose analyzes a smell and transmits the data to a wearable receiver called Scentronic that reproduces it, according to japan.cnet.com.

The team built wristband, handheld and necklace versions of Scentronic, which uses a five-channel piezoelectric atomizer to release the scent. The system relies on Wi-Fi and a cloud server that stores and forwards scent data between devices, and the researchers say it works in real time, with a scent's digital data saved in the cloud for reuse.

In a study, the system was tested with five smells: cake, beer, red wine, coffee and orange. All 16 participants correctly identified the five reproduced scents. Chunyu Yang, a doctoral student at Ohio State University and first author of the Scentronic paper, said the scents were chosen because they are familiar and distinct everyday smells.

Yang said the e-nose combines an array of multiple gas sensors with machine learning to recognize smells, which allows a smaller sensor system suitable for wearables. The paper was published online on October 21 US time.

Earlier scent devices

The researchers suggested possible uses including checking a product's scent before buying online and medical diagnosis. Scent-releasing devices already exist on the market, but they cannot analyze smells, letting users choose only from preloaded scents. In 2014, the oPhone appeared, using scent cartridges called oChips and an iPhone app to release scents called oNotes for 10 seconds, and could reportedly create about 300,000 different smells.