Unveiling crash risks at left-side diverging areas considering drivers’ physiological and behavioral dynamics: a driving simulation study

Diverging areas are widely-recognized as one of the most accident-prone areas on urban expressways. In particular, left-side diverging areas contradict conventional right-side ones in terms of driving habits, driving expectancy, and microscopic maneuvers, leading to sever traffic uncertainty and crash risks. However, to the best of our knowledge, we are still quite ignorant of the characteristics and variations of the driving behaviors at the left-side diverging areas, and the critical factors contributing to crash risks there. To this end, this study aims to investigate crash risks at urban expressway diverging areas and identify contributing factors. To achieve this, a series of driving simulation experiments was conducted, and 3444 traffic conflict cases were identified via the indicator of time to collision (TTC) and recorded along with driving behavioral, physiological characteristics, and personal traits, and the deceleration lane length was also included. A multinomial logit model was employed to explore relationships between identified crash risks and the above factors. The results show that fewer lane changes and longer deceleration lane lengths can effectively mitigate crash risks in the left-side diverging areas. Driving experience, pupil dilation, fixation duration, and saccade duration show significant positive effects on crash risk probabilities in left-side diverging areas. The findings provide insights into the in-depth understanding of the driving behaviors and crash risks at left-side diverging areas, which is supposed to benefit to develop and implement more effective approaches to improve traffic safety of urban expressway.