Researchers have developed an AI-assisted system capable of transmitting smells across the internet. Chunyu Yang and her colleagues constructed an electronic nose and a corresponding smell generator to facilitate this process.
The electronic nose utilizes arrays of sensors composed of metal oxide. An AI algorithm was trained to identify signals from five specific beverages and foods, including coffee, beer, red wine, cake, and orange.
The smell generation component, named Scentronic, uses miniature containers to hold liquids infused with each of the five scents. Electric signals trigger atomisers within the Scentronic system to vibrate and release fine droplets into the air.
Chunyu Yang tested a handheld version of the device on 14 volunteers. Participants correctly identified the transmitted smells an average of 73 per cent of the time.
Yang also created a prototype that uses electromagnetic pumps and a cup of water to reproduce taste. Marianna Obrist at University College London described the concept as "a great idea" but noted, "There’s still a long way to go before we can truly record and replay smell experiences."
Chunyu Yang at the Ohio State University noted that smell remains "still a very undeveloped sense." Yang stated that creating the smell generation device was "the most challenging part" of the project.
