Scentronic System Sends Smells Over Internet With 73% Accuracy in Ohio State University Study

Breakthrough in Digital Scent Technology

Researchers from Ohio State University and Zhejiang Laboratory in China have developed Scentronic, an AI-powered scent system capable of detecting, transmitting, and recreating aromas over the internet.1

1

Published in Science Advances, the technology represents a significant step toward adding smell to digital experiences, addressing what doctoral candidate Chunyu Yang describes as "still a very undeveloped sense" in electronic communication.1

1

The system combines an electronic nose (e-nose) with wearable scent dispensers that can recreate five distinct aromas: coffee, beer, red wine, cake, and orange.2

2

In laboratory tests involving 14 healthy volunteers, participants correctly identified individual smells 73 percent of the time and combinations with approximately 71 percent accuracy.3

3

The system responded to scent signals in 0.6 seconds, though identifying complex food odors required several seconds.3

3

How Machine Learning Powers Scent Detection

The e-nose consists of arrays of metal oxide sensors that detect volatile organic compounds responsible for smells.1

1

When gaseous components like ethanol from alcoholic drinks or limonene from citrus fruits bind to the sensors' surface, this changes their electrical resistance, creating distinctive signals.1

1

Yang and her colleagues trained machine learning algorithms to recognize these patterns and classify them into odor categories.3

3

Training the AI to correctly identify smells proved "the most challenging part" of the study, requiring Yang to expose sensors to various aromas hundreds of times to generate sufficient data.1

1

Wearable Scent Dispensers Recreate Aromas Remotely

Once the e-nose detects and analyzes an odor, it sends scent data over the internet to receiving devices Yang calls Scentronic.2

2

The team created multiple versions including a wristband, handheld model, and necklace.2

2

Each device features a five-channel scent distributor using a piezoelectric atomizer that vibrates in response to electric signals, forcing scented liquid through a cotton filter and releasing fine droplets into the air.1

1

The entire system operates via Wi-Fi and a cloud server that stores and transfers scent data between devices.2

2

However, the technology doesn't transmit actual odor molecules across the internet—instead, it identifies a scent and recreates an approximation using ingredients already loaded into the receiving device.4

4

Applications in Virtual Worlds and Online Shopping

The entertainment industry could be first to develop this digital scent technology, creating more immersive virtual worlds.1

1

Yang suggests users could explore virtual bakeries and smell freshly baked cake, or test fragrances remotely before purchasing.4

4

For online shopping, the technology could let people smell products before buying them, adding a sensory dimension currently missing from e-commerce.2

2

The team also developed a prototype cup with electromagnetic pumps containing flavored liquids that combines scent with taste, allowing users to select a virtual drink through mixed-reality headsets and experience corresponding aromas.1

1

Medical Diagnostics and Healthcare Applications

Beyond entertainment, the Scentronic system shows promise for medical diagnostics. Yang notes the technology could enable doctors to assess distortions in patients' sense of smell related to post-viral illness or neurodegenerative disease.1

1

Loss of smell is a common COVID-19 symptom, while conditions like migraines and pregnancy can cause hyperosmia, a heightened sensitivity to smell.2

2

The system could also make military or astronaut rations more palatable by enhancing the eating experience through scent, since smell transforms how we perceive food.1

1

Challenges and Future Development

Marianna Obrist at University College London called the Scentronic system "a great idea" and a step forward, but emphasized that recognizing a smell and releasing a matching one from a limited library differs significantly from the ease of transmitting sound and vision.1

1

"There's still a long way to go before we can truly record and replay smell experiences," Obrist noted.1

1

The technology cannot yet reconstruct arbitrary smells from basic chemical components, limiting its range and precision.3

3

Researchers plan to test the system outside controlled laboratory conditions, introduce more complex evaluations, and improve controls for potential biases among participants.3

3

Yang hopes to expand the repertoire to create more complex aromas beyond the current five scents.1

1