Abstract
Drowsy driving lead to a serious accident because of low consciousness. This is why driver state estimation method of accurate drowsiness level is required. In previous research, features are calculated from electrocardiogram and breathing while driving. Features are used in pattern recognition to estimate drowsiness level. But this method is unstable in estimated accuracy since value of data is not standard. Our study propose method of estimating driver's drowsiness level by using standard value of heart rate. It is considered that base heart rate during day is minimum heart rate while sleeping. We standardized heart rate during day by base heart rate. As a result, the standard deviation of data improved. We measured eight subjects for three days. After measuring electrocardiogram while sleeping, we measured electrocardiogram and breathing during drowsy driving imitation measurement by using driving simulator. By comparing estimated drowsiness level and true drowsiness level obtained from face image evaluation, we calculated rate of concordance. As a result, estimated accuracy was improved by standardized heart rate while driving using base heart rate.
