Personalized Mixed-Drink Manufacturing System

A personalized mixed-drink system that analyzes preferences using survey and EEG-based learning models to automatically formulate mixing ratios.

This technology is a personalized mixed-drink manufacturing system consisting of a mixing device with separate storage for various liquors, additives, and garnishes, and a control unit that recommends drinks based on a learning model and operates the device. The control unit creates graph patterns of user and liquor characteristics based on survey responses and sends the optimal mixing ratio and garnish selection to the mixing device.

While personalized products are expanding across many fields, custom alcoholic beverages have historically relied on recipes developed by cocktail experts or improvised on the spot, lacking a foundation in individual preference data or scientific analysis. This has limited the ability to meet increasingly segmented personal tastes, create added value through new flavors, and address social demands for reducing alcohol-related harm.

This technology quantifies and uses as learning data the importance users place on psychological factors—such as cognitive ability, emotion, personality, stress, and sociality—as well as physical condition, medical history, drinking context, and liquor attributes like type, ABV, flavor, aroma, color, brand image, and price. It enhances recommendation accuracy by extracting EEG factors (stability, stress, surprise, focus, cognitive load) during surveys and incorporating brain activation patterns captured during consumption, while accumulating user feedback. Applicable to unmanned cocktail dispensers in bars, experiential retail services, and personalized subscription services, it shifts recipe development from bartender intuition to data-driven insights, providing a responsible drinking experience tailored to the user's physical state.

‍

‍

Key Features:

• A liquor mixing device comprising storage units for multiple liquors, additives, and garnishes, which are combined to produce a mixed drink.
• A control unit that executes a recommendation model built on survey-based graph patterns and drink candidates to transmit mixing ratios and garnish selections.
• A control unit that extracts EEG factors—including stability, stress, surprise, focus, and cognitive load—from signals measured while the user completes a survey.
• A process of generating learning data and building a recommendation model by matching individual graph patterns with personalized mixed-drink candidates.

‍

This invention was developed with support from the Ministry of Culture, Sports and Tourism for the development of cognitive-response personalized UI/UX technology for cultural enjoyment by the smart senior generation.

‍