ECG-Based Stress Detection and Productivity Factors Monitoring: The Real-Time Production Factory System.

Area of Science:

• Industrial Engineering
• Occupational Health
• Machine Learning

Background:

• Company success hinges on productivity and quality, influenced by machinery, environment, processes, and human factors like stress.
• Work-related stress is a significant, hard-to-quantify human factor impacting industrial performance.
• Optimizing productivity requires a holistic approach considering multiple simultaneous factors.

Purpose of the Study:

• To develop and validate a system for real-time detection of worker stress and fatigue.
• To integrate worker well-being data with production and environmental monitoring.
• To enable comprehensive analysis for improving productivity and worker sustainability.

Main Methods:

• Utilized wearable sensors for real-time data acquisition.
• Applied machine learning techniques, specifically a 1D Convolutional Neural Network, for stress detection.
• Integrated data streams into a unified platform for multidimensional analysis.

Main Results:

• Demonstrated technical and operational feasibility in an on-field trial.
• Achieved high usability of the developed system.
• Successfully detected stress from ECG signals with 88.4% accuracy and 0.90 F1-score.

Conclusions:

• The system is feasible and usable for real-time stress detection in industrial settings.
• Integrating worker well-being data with operational data enables targeted improvements.
• The approach supports enhanced productivity through sustainable work environments and processes.