True connoisseurs of coffee may soon face serious competition from a new electronic “nose.” According to a study published in the journal IEEE Transactions on Agrifood Elements, this innovative device is capable of determining the aroma of coffee with up to 98% accuracy.
An electronic nose is a device that analyzes gas composition to classify substances. These devices enable real-time monitoring of product quality, evaluation of human health, and support for sustainable agriculture.
Professor Zhong Hong Lee from the National University of Science and Technology in Taiwan explains the main goal of this research: “Our aim is to preserve aromas despite environmental and climate changes, and to ensure the consistency of the quality and taste characteristics of coffee from different regions and crops.”
Coffee flavors are influenced by various factors, including cultivation location, climate, processing methods, and bean genetics. Climate change also plays a significant role in the formation of aroma and taste.
To develop the electronic nose, Professor Lee utilized a set of 8 metal-oxide sensors that are sensitive to specific gases. These sensors collect data on molecules present in gas mixtures, which is then analyzed using artificial intelligence. The researchers trained and tested several AI systems capable of distinguishing between 16 types of coffee from around the world.
All AI algorithms demonstrated impressive results, with accuracy ranging from 81% to 98%, depending on the coffee variety. The best algorithm, a reserve neural network, consistently achieved an accuracy of over 90% and reached 98% for some species.
The data collected by the electronic nose will contribute to the creation of a comprehensive digital library of aromas. This library will record original smells and track how they change over time due to climate factors and growing conditions.
Professor Lee’s team is actively seeking commercialization opportunities for their project and is open to collaboration with enterprises and startups. They also plan to create an electronic language to determine taste, electronic eyes for selecting ripe coffee beans, and a tactile system for assessing bean humidity. “This is a critical step in addressing the recurring challenges faced by farmers,” emphasizes Professor Lee.