Driving fatigue detection based on fusion of EEG and vehicle motion information

Driving fatigue is one of the main causes of traffic accidents. Generally, the single-modal driving fatigue detection methods have low accuracy and weak anti-interference capability in real environment. In this paper, we propose an optimized method for driving fatigue detection that fuses EEG and vehicle motion information. The correlation between EEG signal and vehicle motion information is analyzed. The method extracts the frequency band energy ratio of the four rhythm waves of alpha, beta, theta and delta from EEG. The sample entropy and standard deviation from the steering wheel angle, driving acceleration and vehicle speed is calculated. Pearson correlation analysis is performed on all fatigue characteristic indicators. The strongest correlation combination is selected and input into Support Vector Machine (SVM) to realize the fusion driving fatigue detection. Through experimental verification, we obtained an average classification accuracy of 92.37 %, which is 2.53 % higher than that of EEG driving fatigue detection and 8.78 % higher than that of vehicle motion information-based driving fatigue detection. Multi-source information fusion of driving fatigue detection based on correlation analysis can provide significant insights into how to improve system performance.