Create a free Perfumer & Flavorist account to continue reading

Scientists Put Smell on the Internet, Opening New Possibilities for Digital Fragrance and Flavor

Pixelated retro illustration of a brown and tan computer monitor with geometric pixel art design
Researchers trained an electronic nose to recognize five everyday odors: orange, cake, coffee, beer and red wine.
Mukhamat at Adobe Stock

Digital smell and taste have long been areas of active research and commercialization. To mixed results. 

New research makes a familiar claim: smell could become part of the digital experience. The work describes a system that detects odors, transmits the information over the internet and recreates recognizable scents remotely.

Published in Science Advances, the research introduces Scentronic, an artificial intelligence-assisted olfactory interface developed by researchers from The Ohio State University and Zhejiang Laboratory. The system connects an electronic nose with a wearable scent-delivery device, creating a working framework for remote odor sharing.

The implications extend beyond virtual reality. For the fragrance and flavor industries, the research offers a glimpse of how scent might eventually be integrated into digital commerce, multisensory product experiences and remote sensory evaluation.

Five Scents, One Digital Transmission System

The researchers trained the electronic nose to recognize five everyday odors: orange, cake, coffee, beer and red wine. Machine-learning algorithms classify sensor readings, enabling the system to transmit odor information to a remotely controlled device that releases a corresponding aroma.

In tests with human volunteers, participants correctly identified individual generated odors approximately 73% of the time. Recognition accuracy for pairs of odors was approximately 71%.

The system demonstrated a sensor-to-actuator response time of about 0.6 seconds. However, classifying more complex food odors could take several to tens of seconds.

The distinction is important: the technology demonstrates that selected odor information can be transmitted and converted into recognizable scent cues, rather than that arbitrary smells can already be recorded and reproduced with complete chemical fidelity.

What It Could Mean for Fragrance and Flavor

The work points toward potential applications in virtual and mixed-reality environments, where scent could complement visual and auditory content. Digital shopping experiences could eventually incorporate programmed aromas, while remote sensory research could benefit from more controllable odor delivery.

The researchers also developed a prototype digital cup combining odor stimulation with flavored-liquid delivery. Integrated with a mixed-reality environment, the prototype demonstrated how virtual selections could trigger corresponding physical sensory experiences.

Healthcare is another potential application, particularly remote olfactory assessment and sensory rehabilitation. The researchers also discuss opportunities to enhance food experiences for people facing nutritional challenges or operating in environments where meals may be less appealing.

These remain prospective applications rather than established commercial uses.

The Limitation: Recognizing a Smell Is Not Recreating It Exactly

Scentronic currently works with a predefined selection of odors and preloaded aroma channels. It classifies incoming odor information and activates a corresponding scent, rather than reconstructing the original sample's full chemical composition.

That limitation matters for perfumers and flavorists, whose work depends on complex mixtures, subtle differences and precise sensory fidelity. The system is an early demonstration of digital odor transmission, not a replacement for conventional fragrance creation or a means of transmitting any perfume on demand.

Nevertheless, the research establishes a technological pathway linking chemical sensing, machine learning, wireless communication and programmable odor release.

Source: Chunyu Yang et al., “Scentronic: An artificial intelligence–assisted olfactory interface for remote odor sharing and transmission,” Science Advances (2026). DOI: 10.1126/sciadv.aee9266.

More in Regulatory & Research