Analysis on Characteristics of Dangerous Driving Events via Recorded Data of Drive-Recorder

Abstract

In this research, from the viewpoint of improving traffic safety policies and measures, we focus on clarifying the characteristics of dangerous driving events (stop sign violation, traffic light violation, pedestrian disturbance, excessive speed, etc.) which are thought to be closely related to traffic accidents. Many previous studies on traffic safety have been mainly focused on detecting dangerous road sections which are prone to accidents by using probe data or car accident data to improve road structure. However, according to recent statistics, more fatal traffic accidents have been mainly caused by human factors such as dangerous driving and traffic violations than structural failure or other technical issues of vehicles. Thus, we focus on dangerous driving events or behaviors which can lead to fatal accidents and analyze their occurrence characteristics and correlation with driver attributes. We collected video data and GPS data of drive recorders of 178 residents in Aichi Prefecture to analyze the spatio-temporal characteristics of various types of dangerous driving events and their correlation with driver attributes. In conclusion, we show that the dangerous driving events are correlated with not only driver attributes such as age and gender but also surrounding environments and the driving hours, which are closely related to driver’s psychological state.The results of this research can be expected to contribute to effective driving safety enlightenment campaigns for reducing traffic accidents. We also investigate the relationship between cognitive impairment and dangerous driving of elderly drivers to bring new findings to future traffic safety policies.