Driving Distraction Detection Method Based on Supported Vector Data Descreption Emotion Recognition

摘要

A new approach to recognize the state of driver distraction was proposed. Firstly, the strong relationships between driving distraction and the physiological or visual signals were discussed. Secondly, an algorithm was developed based on support vector data description from the physiological signal databases. And the algorithm consisted of signal preprocessing, feature extraction and pattern classification. To reduce the computational complexity and improve the robustness of the algorithm, the support vector data description was adopted as a pattern classifier. Finally, the results of recognition experiments showed the effectiveness of this method.