An FPGA-Based Machine Learning Tool for In-Situ Food Quality Tracking Using Sensor Fusion

Area of Science:

• Sensor technology and data analysis
• Machine learning applications in food science

Background:

• Sensor arrays are increasingly used in health, bio-signals, and food quality tracking.
• Current machine learning solutions for sensor data analysis are often costly and bulky, limiting in-situ applications.

Purpose of the Study:

• To develop a low-cost, portable machine learning system for food quality assessment.
• To demonstrate the feasibility of implementing sensor fusion algorithms on a single Field Programmable Gate Array (FPGA) chip.

Main Methods:

• Application of machine learning techniques for sensor fusion.
• Implementation of algorithms on a single Field Programmable Gate Array (FPGA) chip.
• Testing the system on an electronic-nose (e-nose) for beef classification and microbial population prediction.

Main Results:

• Successful implementation of machine learning for food quality assessment on an FPGA.
• Demonstrated low-cost, low-power, and compact solution suitable for in-situ deployment.
• Validated the system's capability in beef classification and microbial population prediction.

Conclusions:

• FPGA-based machine learning offers a viable and cost-effective alternative for in-situ food quality analysis.
• The developed system addresses the limitations of traditional bulky computing solutions.
• The approach is applicable to various food manufacturing and quality control scenarios.