AI-powered electronic nose detects a variety of scents for healthcare and environmental applications

The research team has developed the “next-generation AI electronic nose” that distinguishes scents like human olfactory systems and can be analyzed using artificial intelligence. This technology converts fragrance molecules into electrical signals and trains AI models with unique patterns. We have high expectations for personalized healthcare, cosmetics industry and environmental surveillance applications.

This study is published in the Journal ACS Nano. The team is led by Professor Hyuk-Jun Kwon of the Department of Electrical Engineering and Computer Science at DGIST, and holds an integrated master’s and doctoral degrees. Hyuntae, a student as the first author.

Traditional electronic nose (E-North) has already been deployed in areas such as food safety and gas detection in industrial environments, but it has struggled to distinguish between the subtle differences in similar odors and analyze complex scent compositions. For example, distinguishing between flower perfumes with similar notes, or detecting faint odors of fruits approaching decay is challenging for the current system. This gap will increase demand for next-generation e-North technology, increasing accuracy, sensitivity and adaptability.

The researchers were inspired by a biological mechanism known as combinatorial coding. In this mechanism, a single odorant molecule activates multiple olfactory receptors to create a unique pattern of neural signals. By mimicking this principle, the team designed sensors that respond to scent molecules by generating distinct combinations of electrical signals.

AI systems learn these complex signal patterns to accurately recognize and classify different scents, resulting in a high-performance artificial melting platform that surpasses existing technologies.

The new electron nose uses a laser to process thin carbon-based materials (graphene) and incorporates cerium oxide nanocatalysts to create a sensitive sensor array. This single step laser manufacturing method eliminates the need for complex manufacturing equipment and enables highly efficient production of integrated sensor arrays.

In performance testing, the device successfully identifies nine fragrances commonly used in perfumes and cosmetics, identifying accuracy of over 95%. It also allows you to estimate the concentration of each scent, making it suitable for olfactory analysis of fine grains.

This device is extremely thin, flexible, durable and perfect for wearable devices and bright patches attached to your skin and clothing. It can bend more than 30,000 times around a radius of 2.5 mm without degradation in performance.

“A central innovation in our research is the ability to integrate sensors that are sensitive to multiple scents, with the same diverse properties as those of human noses, through a one-stage selective laser manufacturing process,” Professor Kwon said. “We are actively expanding our development and commercialization efforts to apply this technology to the personal healthcare, environmental pollution detection and fragrance industries.”

This study was conducted with a PhD. Student Hyuntae is the first author and corresponding author, Professor Hyuk Jun Kwon.