Classifying driving mental fatigue based on multivariate autoregressive models and kernel learning algorithms

This study developed a driving mental fatigue estimation system based on electroencephalogram (EEG) when he drives a car in a virtual reality (VR)-based dynamic simulator. To classify driver´s mental fatigue status, the features of multichannel electroencephalographic (EEG) signals of frontal, central and occipital are extracted by multivariate autoregressive (MVAR) model. Then kernel principal component analysis (KPCA) and support vector machines (SVM) are combined to identify three-levels driving mental fatigue. The results show that KPCA is an good feature extractor which can effectively reduce the dimensionality of the feature vectors. The KPCA-SVM shows good performance with higher classification accuracy (81.64%) across 10 subjects. This method could be an potential approach of classification of driving mental fatigue.