Institute of Transportation, MOTC

       Due to the growing trend of public transportation, inter-city bus (highway bus) as ground transport plays an essential role in Taiwan. However, the traffic accidents involving with bus usually accompany serious casualties and financial loss. The fatality rate of bus are 10 times higher than sedans in recent years, and it shows urgency of bus safety in Taiwan. Thus, driving behavior is gradually valued by interbus carriers. The bus driver is charged with serious responsibility, and his driving behavior is influenced to a significant degree by his own human factors. This study will collect information of inter-city bus driver human factors and the status of aberrant driving behavior from the case study company. A relative risk level evaluation mechanism will be developed based on the frequency and distribution of aberrant driving behavior. The research aims to enhance efficiency of the fleet management system and highway safety in general through quantifying relative driving risk of each driver. We apply artificial neural network (ANN) models to predict the frequency of aberrant driving behavior and the risk level of each driver by individual human factors. The predictive models perform high accuracy in case. Spearman correlation coefficient was used to calculate the correlation between the human factors and driving risk. Driving fatigue, symptom, disease and high neuroticism would cause high driving risk; Enough sleep hours, high agreeableness and high annual household income lead to low driving risk. By establishing a systematic driving risk assessment mechanism, inter-city bus industry can reduce the occasion of traffic accidents and further raise corporate integrity and reputation.