Towards analyzing crash events for novice drivers under reduced-visibility settings: A simulator study

Abstract

Road accidents are one of the primary concerns and critical issues that encounters societies nowadays: Crash events analysis is a key role in improving traffic safety and reducing potential inconveniences to road users. As such, novice drivers continue to be overrepresented in fatalities and injuries arising from crashes, especially in those that occur during nigh times under rainy weather conditions. In this study, we aim to investigate road crash events for novice drivers under reduced-visibility scenarios during multiple night-time driving simulations that have been conducted using a desktop driving simulator. This paper depicted the effect of both light rain and heavy rain on traffic safety by endorsing real-time driver inputs established as throttle pedal position, brake pedal position and wheel angle. To the authors' knowledge, minimal work has been directed to the examination of light and heavy rain on novice drivers crash events based on driver inputs during night times. Artificial Neural Networks (ANN) and Decision Trees (DT) machine learning models have been developed to analyze crash events; results depict that ANN model exhibited the best performances in terms of accuracy and AUC measures during all-weather covariates. Conventionally, and based on the findings, new insights into night-related crash events' assessments for novice drivers could be harnessed to assist enforcement endeavors to design crash avoidance/warning systems under reduced-visibility settings.