- Researchers have developed a wearable device capable of recreating food aromas.
- Participants achieved a 73% accuracy rate in identifying a single aroma.
- The new system employs machine learning to identify scents and release fragrances.
- The technology currently relies on pre-loaded aromas, limiting its range of applications.
An experimental system aims to incorporate the aroma of coffee into virtual gatherings, connecting scent sensors to a wearable fragrance release device via the internet. Researchers from Ohio State University in the United States and a laboratory in Zhejiang, China, developed this automated system, which uses artificial intelligence to identify scents and trigger their release near another person's nose. According to a report in Science Advances, this technology combines chemical sensing, machine learning, and wearable devices presented in the form of bracelets or necklaces.
In laboratory tests involving 14 healthy volunteers, participants were able to identify a single aroma with an accuracy rate of 73%, while the accuracy rate for identifying combinations of aromas was approximately 71%. This finding marks a significant step toward adding scent to digital experiences, although the system currently relies on pre-loaded fragrances.
The challenge of reconstructing scents differs from that of recreating images or sounds. Cameras and microphones capture light and sound waves, while scents originate from chemical compounds that must be detected and actually released. Previous attempts included bulky desktop devices, scratch-and-sniff cards, and virtual reality headsets. Some methods even required the receiving personnel to manually release fragrances.
Technical Challenges in Scent Recognition
This new system uses machine learning algorithms to identify patterns associated with volatile organic compounds, the molecules that are the source of scents, and classifies them into scent categories. This information is transmitted over the internet to a receiving device called Scentronic, which releases the corresponding fragrance near the user's nose. This process involves sending digital commands between devices, while the scent itself comes from materials pre-loaded into the wearable device.
Researchers tested five scents: beer, cake, coffee, orange, and red wine. Volunteers were exposed to these scents both individually and in combinations, including beer with cake and coffee with orange. Participants achieved a 73% accuracy rate when identifying a single aroma. The accuracy rate for identifying combined scents was close to this level, reaching approximately 71%. Speed was also a key focus, with the system responding to scent signals within 0.6 seconds; however, identifying more complex food aromas took several seconds.
This distinction is crucial, as reacting to a signal and accurately classifying an unfamiliar chemical mixture are two different tasks. The research team also developed a version that can release aromas infused with flavor liquids, extending the concept from remotely sniffing objects to reconstructing certain aspects of the drinking experience. In potential virtual reality applications, users could select a digital beverage and receive the corresponding flavor liquid at home.
Future Applications and Challenges
The technology currently cannot reconstruct arbitrary scents from their basic chemical components but selects from pre-loaded scent channels, limiting the range and accuracy of what it can reproduce. Researchers plan to test the system outside of controlled laboratory conditions, conduct more complex evaluations, and improve control over potential biases among participants. Possible applications include virtual social gatherings, online shopping, and medical research involving sensory perception or loss. Current work demonstrates a limited but automated connection between detecting scents and reconstructing corresponding fragrances elsewhere, and expanding this connection to the various scents people encounter daily remains the next challenge.
Potential Impact of Wearable Scent Devices
The development of this technology marks a further expansion of digital experiences, particularly in virtual social settings, where the recreation of scents can enhance user immersion. Although the current system still relies on pre-loaded aromas, overcoming this limitation in the future could broaden its applications, including online shopping and medical research. As technology advances, the ability to reconstruct various scents may change the way people interact and enhance sensory experiences.

