Abstract
Drowsy driving is known to raise the risk of accidents. Therefore, many researcher estimate the driver's drowsy state using vehicle operation and biological signals. In our previous study, we proposed a method to predict the time of driver's drowsy state using features before driving. However, we could not reveal why prediction accuracy improved. In this study, we revealed that there are relationships between driver's state before driving and subsequent vehicle operation. After receiving informed consent from 6 subjects, they were subjected to the following experiments. To evaluate driver's state before driving, we measured subjective evaluation and biological signals. Then, we measured vehicle operation using driving simulator. We extracted driving pattern near the intersection and the traffic ‘stop sign’ from acceleration and deceleration. Then, we compared the driver's state before driving and driving pattern. As a result, when driver is different from the usual state, the probability of different driving pattern from usual driving increased. Therefore, our results showed possibility of having a relationship between driver's state before driving and subsequent vehicle operation.
