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

Real-time monitoring of operator fatigue is critical for ensuring operational safety and reliability in dynamic industrial environments, especially under swing conditions such as offshore floating nuclear power platforms. This study proposes a multimodal fatigue monitoring framework based on surface electromyography (sEMG) and heart rate variability (HRV) sensors for real-time fatigue recognition. Experiments were conducted on a six-degree-of-freedom motion platform with three swing levels, involving 23 participants performing simulated emergency operation tasks. Four machine learning models (Naive Bayes, K-Nearest Neighbor, Multilayer Perceptron, and Random Forest) were employed for fatigue state classification. The results show that the Random Forest model achieves the best performance, with an overall accuracy of 98.2%, 100% true positive rate for the normal state and fatigue, and 66.7% true precision for severe fatigue. The proposed multimodal fusion method effectively suppresses motion artifacts and improves recognition robustness under swing interference. Rigorous subject-level stratified cross-validation eliminates sample leakage risks; bootstrap confidence intervals and pairwise significance tests statistically verify model performance differences; class imbalance mitigation strategies are deployed to quantify uncertainty for the scarce severe-fatigue category; literature-supported Borg CR-10 grading thresholds are validated via retrospective cutoff sensitivity analysis to guarantee reliable fatigue labeling. This sensor-based intelligent monitoring system provides a reliable solution for real-time fatigue detection of operators in dynamic digital industrial scenarios, supporting accident prevention and sustainable operation of high-risk industrial systems.