Perceived fatigue progression tracking during manual handling tasks using sEMG recordings.

Area of Science:

• Occupational Health
• Biomedical Engineering
• Ergonomics

Background:

• Physical fatigue is a major cause of work-related musculoskeletal disorders.
• Understanding fatigue during manual handling is crucial for prevention.
• Myoelectric manifestation of fatigue (MMF) indicators are used to assess physical exertion.

Purpose of the Study:

• To correlate changes in MMF indicators with perceived exertion during manual handling.
• To compare the effectiveness of linear and complexity-based MMF indicators.
• To develop a machine learning model for fatigue classification.

Main Methods:

• Surface electromyography (sEMG) and inertial measurement units were used.
• sEMG recordings were segmented by activity and joint range of motion.
• Linear and complexity-based MMF indicators were extracted and correlated with perceived exertion.

Main Results:

• Complexity-based MMF indicators showed significant correlations with perceived fatigue in most muscles studied.
• Linear indicators showed significant correlations primarily in lower leg muscles.
• A deep learning model achieved 69% accuracy in classifying five fatigue stages.

Conclusions:

• Complexity-based MMF indicators are superior to linear indicators for monitoring perceived fatigue.
• MMF indicators can be utilized with machine learning for practical fatigue classification.
• This approach offers potential for personalized fatigue and health monitoring in occupational settings.