IEEE - Inst Electrical Electronics Engineers Inc Acta Medica Okayama 2169-3536 10 2022 Sensitivity of PERCLOS70 to Drowsiness Level: Effectiveness of PERCLOS70 to Prevent Crashes Caused by Drowsiness 70806 70814 EN Atsuo Murata Department of Intelligent Mechanical Systems, Graduate School of Natural Science and Technology, Okayama University Toshihisa Doi Department of Intelligent Mechanical Systems, Graduate School of Natural Science and Technology, Okayama University Waldemar Karwowski Department of Engineering and Management Systems, University of Central Florida It has been reported that many crashes are caused by drowsiness. Thus, it is critical to predict the occurrence of severe drowsiness that may result in a crash by means of an effective measure. The aim of this study was to investigate whether percentage closure (PERCLOS) of 70% was useful for evaluating drowsiness level of individual drivers and preventing crashes caused by drowsy driving using a driving simulator system. The first experiment measured PERCLOS70 during both aroused and drowsy states in a driving simulator task and investigated how PERCLOS70 changes when a participant fell asleep. In the second experiment, we measured PERCLOS70 and investigated the relation between PERCLOS70 and Karolinska Sleepiness Scale (KSS) ratings during a simulated driving task. The aggregated mean PERCLOS70 was significantly higher when participants fell asleep than when they were aroused. This tendency was also observed for individual participants. The aggregated mean PERCLOS70 was found to be sensitive to changes in KSS scores and increased with increasing KSS score. Linear trend analysis revealed a significant increasing trend for PERCLOS70 as a function of the KSS rating. This tendency was also observed for individual participants. PERCLOS70 was found to be sensitive to the drowsiness level both for data aggregated across all participants and data for individual participants. The main findings of the two experiments reported herein suggest that PERCLOS70 can be used effectively to evaluate drowsiness of individual drivers and prevent crashes caused by drowsy driving. No potential conflict of interest relevant to this article was reported.