Skip to Main Content
A novel EEG-based system for driver's sleepiness detection is proposed. Driver's sleepiness is an important factor in many accidents. Therefore, real-time sleepiness detection can restrain accidents effectively. In this study, SSVEPs are used for running the proposed system. In order to generate SSVEPs in the brain activities, two experimental setups consisting four single and paired LEDs are proposed. In addition, the effect of two different FFT-based feature extraction methods, and two different classifiers of the LDA and the SVM on the accuracy of the system are studied. Related features are extracted from three different segments (sweep lengths) of 0.5, 1, and 2 seconds. The experimental results show that higher sweep lengths have higher accuracies and the SVM classifier, experimental setup of 4-paired LEDs and sweep length of 1 second has the highest ITR value of 24 bits/min. Therefore, this study demonstrates the feasibility of the proposed system in a practical driving application.