AN ANALYSIS OF TRAFFIC ACCIDENT RISKS CONSIDERING LONG CONTINUOUS DRIVING VEHICLES

This study analyses the impact of Long Continuous Driving Vehicles (LCDV) on road traffic accident risks. Traffic accidents are a big concern of society in many countries, because they often lead to non-recurrent road congestion that significantly degrades reliability in travel time and results in huge economic loss. Majority of recent traffic accidents are caused by careless driving, which is assumed to result from long-continuous driving. However, there is no scientific evidence that relates the driving hours and traffic accident risks. This study aims to evaluate the impact of LCDVs on traffic accident risks along a highway section in Ehime prefecture, Japan. An index is introduced to represent the impact of LCDVs, the proportion of LCDV, at each road section and time interval. The index is calculated using detailed probe vehicle data from commercial vehicles. Then, a Poisson regression model is developed to estimate an accident risk, expected number of accidents per vehicle kilometre travelled, depending on various explanatory variables; the variables include road geometry types, temporal differences and the proportion of LCDV. The log-likelihood ratio of the estimated model exceeds 0.2, confirming that the model fits enough to the data. The coefficient of the proportion of LCDV is significantly positive, which shows that the existence of LCDVs in traffic stream increases accident risks. The result can be utilised to design recommended driving hours and to assess benefit of taking a rest while driving.