A Versatile, Machine-Learning-Enhanced RF Spectral Sensor for Developing a Trunk Hydration Monitoring System in Smart...

Area of Science:

• Microwave Engineering
• Sensor Technology
• Smart Agriculture

Background:

• Automated radio frequency (RF) scattered parameter acquisition is crucial for non-invasive monitoring.
• Existing systems often lack compactness and standalone capabilities for field deployment.
• Integration of computational platforms enhances data processing and analytical potential.

Purpose of the Study:

• To develop a standalone, compact microwave sensing system for automated S-parameter acquisition.
• To enable non-invasive monitoring of external matter and environmental variables.
• To demonstrate the system's capability for real-time analysis using embedded machine learning.

Main Methods:

• Integration of a NanoVNA and Raspberry Pi Zero W for S-parameter (50 kHz-4.4 GHz) recording.
• Implementation of dual recording modes (manual and automatic) powered by a single battery.
• Embedding machine learning algorithms within the Linux-based system architecture.

Main Results:

• Successful development of a compact, battery-operated microwave sensing system.
• Demonstrated capability for automated S-parameter data collection.
• Validated the system's potential for greenwood hydration detection using an RF patch antenna and ML analysis.

Conclusions:

• The developed system offers a flexible and portable solution for RF scattered parameter acquisition.
• Embedded ML algorithms facilitate automated, real-time analysis of environmental variables.
• The system shows significant promise for applications in smart agriculture and non-invasive material monitoring.