Intelligent Multigas Monitoring: A Reconfigurable RFID Sensor with Machine Learning-Assisted Decoding for O₂ and CO₂

Area of Science:

• Sensor Technology
• Machine Learning
• Chemical Engineering

Background:

• Traditional fermentation monitoring faces limitations in precision and real-time data acquisition.
• Existing wireless sensors often struggle with dual-parameter detection and complex environmental conditions.
• Need for advanced sensing solutions to support digital transformation in the brewing industry.

Purpose of the Study:

• To develop a reconfigurable antenna-based RFID sensor for simultaneous wireless monitoring of O2 and CO2.
• To integrate a random forest algorithm for accurate, high-precision gas concentration inversion.
• To demonstrate the application of this technology in fermentation environments, specifically wine production.

Main Methods:

• Design of a reconfigurable RFID tag antenna using complementary split-ring resonators (CSRR) and photodiodes for enhanced coding capacity.
• Utilization of SnS2/ZnO/NiO and SnO2/CuO/TiO2 nanocomposites as gas-sensitive layers for improved adsorption and conductivity.
• Development of a multidimensional RF feature-decoding framework based on random forest regression for nonlinear signal inversion.

Main Results:

• The sensor achieved high-precision detection of O2 (1000-250,000 ppm) and CO2 (500-50,000 ppm) with rapid response times.
• Demonstrated excellent repeatability, long-term stability, and significant amplitude changes (14.79 dB for O2, 17.13 dB for CO2).
• Random forest model exhibited high accuracy and generalization ability, outperforming traditional linear fitting methods in gas concentration prediction.

Conclusions:

• The developed RFID sensor system enables accurate, wireless, dual-parameter gas monitoring in fermentation.
• The integration of reconfigurable antennas and random forest algorithms overcomes previous accuracy limitations.
• This technology supports the digital transformation of the brewing industry, improving product quality and reducing costs.