Driver drowsiness detection based on classification of surface electromyography features in a driving simulator.

Area of Science:

• Biomedical Engineering
• Neuroscience
• Transportation Safety

Background:

• Driver drowsiness is a major global cause of fatal road accidents.
• Objective detection of drowsiness is crucial for enhancing road safety.

Purpose of the Study:

• To detect driver drowsiness by analyzing surface electromyography (sEMG) signal features.
• To evaluate the effectiveness of different classifiers for drowsiness detection.

Main Methods:

• sEMG signals were recorded from upper arm and shoulder muscles of 13 healthy subjects in a driving simulator.
• Features like range, variance, relative spectral power, kurtosis, and shape factor were extracted from 30-s epochs.
• Six classifiers were applied, with drowsiness levels assessed using the Observer Rating of Drowsiness scale.

Main Results:

• The k-nearest neighbor (KNN) classifier demonstrated high performance in drowsiness detection.
• KNN achieved 90% accuracy, 82% precision, 77% sensitivity, and 92% specificity.
• A binormal function was fitted for each extracted feature to aid classification.

Conclusions:

• sEMG signal feature analysis is a viable method for detecting driver drowsiness.
• The KNN classifier shows significant potential for real-time drowsiness detection systems.
• This research contributes to developing technologies for preventing fatigue-related traffic accidents.