A Novel Drinking Category Detection Method Based on Wireless Signals and Artificial Neural Network.

Area of Science:

• Consumer Science
• Signal Processing
• Machine Learning

Background:

• Growing consumer demand for healthier beverage options like Not From Concentrated (NFC) juice and packaged water.
• Limitations of current beverage detection systems, including reliance on specialized equipment and uncommon signals.
• Potential for automated systems like vending machines to benefit from improved beverage classification.

Purpose of the Study:

• To develop a novel and accessible method for detecting beverage categories.
• To leverage widely available WiFi signals for beverage identification.
• To improve the accuracy and efficiency of automated beverage detection systems.

Main Methods:

• Utilizing WiFi signals to detect different beverage categories.
• Capturing WiFi signals using wireless devices and extracting Channel State Information (CSI).
• Applying artificial neural networks (ANN) for classification after noise removal and feature extraction.

Main Results:

• Demonstrated high accuracy in detecting various beverage categories.
• Successfully leveraged distinct multipath propagation patterns of WiFi signals caused by different beverages.
• Validated the effectiveness of the ANN model in classifying beverages based on CSI data.

Conclusions:

• The proposed method offers a promising, accurate, and practical solution for beverage category detection.
• WiFi-based sensing presents a viable alternative to traditional methods, utilizing ubiquitous technology.
• This technology can enhance consumer experiences and operational efficiency in retail and vending environments.