Real-Time Fatigue Monitoring Using sEMG and HRV Sensors for Industrial Operators Under Swing Conditions

Area of Science:

• Industrial Safety and Human Factors Engineering
• Biomedical Signal Processing
• Machine Learning in Health

Background:

• Operator fatigue poses significant risks in dynamic industrial environments, particularly on offshore platforms.
• Real-time monitoring is essential for preventing accidents and ensuring operational reliability.
• Existing methods struggle with motion artifacts and interference in dynamic conditions.

Purpose of the Study:

• To propose and validate a multimodal fatigue monitoring framework using sEMG and HRV.
• To assess the framework's effectiveness under simulated swing conditions.
• To develop a robust, real-time fatigue recognition system for industrial operators.

Main Methods:

• A multimodal framework integrating sEMG and HRV sensors was developed.
• Experiments involved 23 participants on a motion platform simulating offshore platform swings.
• Four machine learning models (Naive Bayes, KNN, MLP, Random Forest) were evaluated for fatigue classification.

Main Results:

• The Random Forest model achieved 98.2% overall accuracy in fatigue detection.
• The multimodal approach demonstrated robustness against motion artifacts and swing interference.
• High true positive rates were observed for normal and fatigue states, with notable precision for severe fatigue.

Conclusions:

• The proposed sensor-based intelligent monitoring system offers a reliable solution for real-time operator fatigue detection.
• This system enhances safety and supports sustainable operations in high-risk digital industrial scenarios.
• The multimodal fusion effectively improves recognition accuracy and robustness in challenging environments.