Research on Recognition Method of Driving Fatigue State Based on Sample Entropy and Kernel Principal Component...

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Driving fatigue recognition is crucial for road safety.
• Existing electroencephalography (EEG)-based methods have limitations in accuracy due to signal nonlinearity.
• Nonlinear characteristics of EEG signals pose challenges for accurate driving fatigue detection.

Purpose of the Study:

• To propose an improved driving fatigue recognition method using EEG signals.
• To address the limitations of current methods by combining sample entropy and kernel principal component analysis.
• To enhance the accuracy and robustness of driving fatigue detection.

Main Methods:

• Utilized sample entropy (SE) for its high recognition accuracy.
• Employed kernel principal component analysis (KPCA) for nonlinear dimensionality reduction.
• Integrated SE and KPCA, termed SE_KPCA, with a support vector machine (SVM) classifier.
• Compared SE_KPCA against methods using fuzzy entropy (FE) and combination entropy (CE) with KPCA.

Main Results:

• The proposed SE_KPCA method demonstrated effective driving fatigue recognition.
• Experimental results confirmed the efficacy of the combined SE and KPCA approach.
• The method showed superior performance compared to other entropy-based methods when combined with KPCA.

Conclusions:

• The SE_KPCA method offers a promising approach for accurate driving fatigue detection.
• Combining SE with KPCA effectively handles the nonlinear nature of EEG signals for fatigue recognition.
• This approach represents a significant advancement in EEG-based driving fatigue monitoring systems.