A noninvasive real-time solution for driving fatigue detection based on left prefrontal EEG and eye blink

Abstract

According to the portable and real-time problems on the driving fatigue prevention based on EEG, a headband integrated with Thinkgear EEG chip, tri-axial accelerometer, gyroscope and Bluetooth is developed to collect the subject’s blink, the Attention and Meditation of the left prefrontal EEG. The comparison between Attention and Meditation of the left prefrontal EEG is discussed at first when the subject is in the state of concentration, relaxation, fatigue and sleep. The slide window and k-NN algorithm are introduced to develop a new method for driving fatigue detection based on subject’s blink and the correlation coefficient between Attention and Meditation. Lastly, a software running on a smart device is developed based on above technologies, it can issue alarm and play music when it detects driving fatigue. The experiment proves that it has noninvasive and real-time advantages, while its sensitivity and specificity are 73.8 % and 88.6 % respectively.