Research on time relevant variables based fatigue level prediction model

Time based regulation is considered an effective way to reduce the driver fatigue risk. However, the maximum driving hours or limitation of accumulated driving hours differs from countries. The purpose of this study is to estimate the time-based fatigue level and establish the prediction model. Twenty-seven field driving trails were conducted to gather the fatigued driving data. Then the fatigue level is measured by Karolinska Sleepiness Scale: KSS. Finally, one linear regression model considering the variable of time on task, sleep hours and break hours was proposed. Validation analysis of model indicated that the correlation coefficient between KSS and predicted fatigue level is 0.86. The prediction model regarding the fatigue level could be a good reference for optimizing working schedule for the commercial transportation companies. The private drivers could also self-evaluate their fatigue driving risk as well.