Monitoring attentional level of pilots can reduce the number of accidents caused by low attentional level during driving tasks. in this project, we evaluate driver's level of attention based on electroencephalogram (EEG) data.
We established machine learning classifiers (Logistic Regression, SVM, Gbdt classifiers, etc.) and a CNN model to identify human EEG signals in different states. The CNN model takes raw EEG segments as input, while other classifiers take extracted features in the frequency domain as well as sample entropy values as input.
In results, we compared the accuracy and robustness of different models using EEG data from multiple human subjects.
