Scentronic technology lets you smell the Internet

Have you ever bought a perfume online with great reviews and amazing packaging, only for it to arrive and the first smell makes you cry… tears of love? grief? Well, that could soon be a thing of the past, as scientists have discovered a way to transmit smells over the Internet. Technically speaking.

Researchers at Ohio State University and the Zhejiang Laboratory in China have created a system that detects an odor in one location and reproduces the corresponding aroma in another location.

The system consists of two devices: an electronic nose (e-nose) that detects chemicals in the air at the sending end and a device called Scentronic that releases fragrances at the receiving end. Machine learning software and the Internet connect them. So far, the team has demonstrated the system with beer, pastries, coffee, oranges and red wine.

We discovered a long time ago how to capture and transmit images and sounds. Specialized devices can even reproduce aspects of touch. Smells, however, have proven more complicated. A camera records light falling on a sensor, while a microphone records changes in air pressure caused by sound waves. Both light and sound follow the laws of physics, a framework that works perfectly with electronics, allowing them to be replicated digitally.

A smell, however, is chemical and is made up of molecules that float in the air, often in complex mixtures. Their identities and concentrations determine what you perceive. A scent has no direct equivalent to a photograph or audio file. So how do you capture one and how do you convert what you’ve captured into digital signals? These are the questions the researchers set out to answer.

The electronic nose is in the shape of a circular circuit board carrying six metal oxide semiconductor gas sensors. Its main sensor material is tin dioxide.

In normal air, oxygen molecules deposit on the sensor surfaces and pull electrons away from the material, increasing its electrical resistance. When volatile organic compounds (VOCs) from food or beverages reach these surfaces, some react with the adsorbed oxygen and return electrons to the tin dioxide. The resistance drops and the circuit registers a corresponding specific increase in voltage.

Ethanol from alcoholic beverages, limonene from citrus fruits, and diacetyl from buttery aromas are some of the VOCs that can cause these reactions. Because each sensor responds differently, each source produces a distinctive voltage pattern across the entire array. Think of it as a chemical fingerprint, albeit a fairly blurry one.

Now, data is available that loosely relates an odor to a voltage pattern. However, precise matching is necessary for the system to be useful. To do this, the researchers turned to AI.

They collected 250 labeled recordings, 50 for each of the five foods and drinks, and measured the changing voltages of the sensors for 200 seconds per recording. They then trained and compared several machine learning models to associate those patterns with the correct categories. Once the system recognizes a familiar pattern, it can assign a label such as “brown” or “orange.” A computer sorts the digitized signals and then transmits the result through a cloud server connected to the Internet.

Now we go to the receiving end, where the Scentronic device is located. The device contains miniature chambers filled with pre-prepared aromatic liquids, each connected to its own release channel. A cotton filter draws liquid into a piezoelectric atomizer, a small disk that flexes when electricity is applied. Driven at approximately 110,000 vibrations per second, it forces liquid through microscopic openings in the form of a mist of fragrant droplets.

When a “coffee” instruction arrives as electronic signals, the receiver activates the channel loaded with a coffee-like fragrance. An ESP32 microcontroller controls the atomizer with rapid electrical pulses, adjusting the spray frequency to regulate the amount of aroma released. You can also trigger multiple channels together to create mixes, varying their relative contributions by controlling each channel independently.

It is important to note that the Scentronic device does not create the scent from scratch. It already comes preloaded with the five possible accents. The AI ​​identifies the odor category, while the receiver follows programmed instructions to release the corresponding fragrances already loaded inside.

The team explored several Scentronic designs, including a lightweight bracelet with two aroma channels, a necklace with five arranged like pearls on a string, and a wearable version with five channels. The bracelet can be raised towards the nose for sniffing, while the collar keeps the emitters close to the face.

In controlled tests, the e-nose’s machine learning models classified the five odor categories with accuracies ranging from 96% to 100%, depending on the method used. However, when 14 volunteers identified fragrances from the portable Scentronic, they were about 73% correct for individual scents and 71% correct for combinations of two scents. In a separate intensity rating test, accuracy was around 59%.

It should be noted that wearables that generate odors do They already exist. In 2023, for example, researchers in Hong Kong demonstrated wireless virtual reality interfaces that released pre-loaded fragrances, plus we’ve seen a face-mounted version capable of presenting more than a dozen scents. However, Scentronic’s equipment combines electronic detection, artificial intelligence recognition, Internet communication and odor release into a remote exchange system.

Of course, there are some big catches. Currently, the e-nose recognizes only five categories and only one product from each was used in the training. Scentronic also needs a suitable fragrance pre-loaded for each odor it is expected to produce. You can’t smell an unknown perfume, discover its full molecular recipe, and faithfully recreate it for an online buyer. With only five channels, the result is a recognizable substitute for the original smell.

Still, there is room for that to be useful. One day, online stores could allow customers to try scents before ordering, although the right fragrances would have to be available at the receiving end. A more practical application may be virtual reality, where a game can ship with a dedicated Scentronic scent package that is released depending on the game. Although (again), the electronic nose does not necessarily have a role in this scenario.

The researchers even added flavor to the matter.

They built an experimental cup with five aroma channels and five miniature electromagnetic pumps that drip flavoring liquids into the water. When someone wearing a mixed reality headset selects a virtual beer or coffee, the cup releases a corresponding aroma and adds the corresponding liquid for flavor. They also demonstrated responses to beer, coffee, and their combination. The prototype shows how sights, sounds, smells and tastes could work together.

“This has broad applications for nutritionally challenged patients and populations such as astronauts (facing unattractive dietary supplements) or warfighters (consuming field rations), where improving perceived food quality can have critical nutritional and performance outcomes,” the researchers wrote. Although (one last time), couldn’t they just add the improvements themselves? It’s not exactly easy to identify a scenario that fits the system as it is. That said, is a prototype that can be developed for many future applications.

The next step for researchers is more sensitive sensors, larger training data sets, additional fragrance channels, and better handling of complex mixtures. They also want more rigorous human testing and standardized methods for evaluating medical odors. For now, those regrettable blind perfume purchases will have to continue.

The study is detailed in the journal. Scientific advances.

Source: American Association for the Advancement of Science via EurekAlert