Super wavelet for sEMG signal extraction during dynamic fatiguing contractions

Area of Science:

• Biomedical Engineering
• Signal Processing
• Kinesiology

Background:

• Muscle fatigue detection is crucial for preventing injuries and optimizing training.
• Surface electromyography (sEMG) is a common method for assessing muscle activity.
• Existing wavelet functions may not optimally capture the nuances of fatiguing sEMG signals.

Purpose of the Study:

• To develop and validate a new algorithm for classifying muscle fatigue from dynamic contractions.
• To utilize a genetic algorithm (GA) for evolving a pseudo-wavelet function tailored for sEMG decomposition.
• To enhance the accuracy of muscle fatigue classification compared to standard wavelet methods.

Main Methods:

• Recorded sEMG signals during fatiguing dynamic contractions of the biceps brachii from thirteen subjects.
• Employed a genetic algorithm (GA) to evolve a pseudo-wavelet function for optimal sEMG signal decomposition.
• Trained and tested the algorithm using labelled 'Fatigue' and 'Non-Fatigue' sEMG trials, with 70% for tuning and 30% for testing.

Main Results:

• The evolved pseudo-wavelet function significantly improved muscle fatigue classification rates by 4.45 to 14.95 percentage points (p<0.05) compared to standard wavelets.
• Achieved an average correct classification rate of 87.90% for muscle fatigue.
• Demonstrated the effectiveness of the GA-evolved pseudo-wavelet in decomposing and classifying sEMG signals.

Conclusions:

• The developed GA-based pseudo-wavelet algorithm offers a superior method for classifying muscle fatigue from sEMG data.
• This approach provides a more accurate and reliable tool for assessing muscle fatigue in dynamic contractions.
• The findings have implications for sports science, rehabilitation, and ergonomic assessments.