AI-driven 5G IoT e-nose for whiskey classification

Area of Science:

• Artificial Intelligence
• Chemical Sensing
• Machine Learning

Background:

• Quality control in whiskey production is crucial for final product assessment.
• Distillery processes generate byproducts like acetone that require monitoring.
• Electronic noses offer a non-destructive method for chemical compound analysis.

Purpose of the Study:

• To design, implement, and validate an AI-driven electronic nose for classifying whiskey and acetone.
• To differentiate between whiskey and acetone, and discriminate between three types of whiskey.
• To enhance quality control in whiskey production through accurate odor classification.

Main Methods:

• Utilized an electronic nose based on arrays of single-walled carbon nanotubes.
• Investigated various strategies for classifying odor data.
• Employed a random forest algorithm for data analysis and classification.

Main Results:

• Achieved 99% accuracy in classifying whiskey and acetone, with inference times under 1.8 seconds.
• Demonstrated high accuracy (around 96%) in distinguishing between different whiskey types.
• Successfully validated the AI-driven electronic nose architecture for its intended applications.

Conclusions:

• The developed AI-driven electronic nose architecture is highly effective for classifying whiskey and acetone.
• The system provides a reliable tool for quality prediction in whiskey production.
• The random forest approach offers a robust and efficient method for odor data classification.