入眠予兆現象捕捉のための呼吸に関する研究

抄録

In Japan, drowsy driving is one of the main causes of traffic accidents. It is highly likely that drowsy driving leads to a serious accident, because we are not able to take an action to avoid danger if we fall asleep during driving. To prevent traffic accidents caused by drowsy driving, there are many researches aiming to estimate drivers' sleepiness by using their physiological signals. In this paper, we focused on respiration and pulse wave as physiological signals for sleepiness estimation, because they are easily measured without interfering driving environment by a sensor-equipped automobile seat. As an index calculated from pulse wave, RR interval (an interval of heart beat) has been often used to estimate sleepiness. However, it is not greatly known that which respiratory index is effective for sleepiness estimation. Based on the consideration above, this paper aims to clarify the respiratory index greatly correlated with sleepiness, and then, to estimate sleepiness more accurately making use of respiration and pulse wave information. In our experiments, we measured respiration and pulse wave simultaneously with evaluating sleepiness from facial expression of the subject. Through the experiments, it was found that inspiratory time, expiratory time and expiratory volume have correlation with sleepiness and in particular expiratory time has the most highly correlated with sleepiness. Moreover, using these three respiratory indexes together with RR interval, we estimated sleepiness by the method of multiple regression analysis. As a result, we demonstrated more accurate sleepiness estimation than the previous method in eighty percent of the experiments.