A Method For Highway Collision Rate Estimation Considering Traffic Flow State

Abstract

Among traffic-related variables, vehicle speed is thought to be one of the major variables closely related to collision occurrence. This study aims to develop a methodology to estimate collision rate considering the traffic state in a freeway section focusing on the speed variables. To define the traffic condition of a section, four section-based traffic states were used according to the speed of upstream and downstream locations. The collisions were classified into three types (Rear-end, Sideswipe, and Others), and two variables (speed difference between upstream and downstream and average speed of upstream and downstream) were considered as independent variables for standard linear regression analysis. The findings showed that both speed difference and average speed had an influencing explanation to state the collision rate by each type of collision and traffic state. Rear-end and sideswipe collisions showed different collision rate pattern with the others, and the influence of the speed difference between upstream and downstream changed according to the traffic state. Through regression analysis, rear-end and sideswipe collision rates were expressed by the speed difference, while other types of collisions were described by the average speed. The research can provide a methodology to understand the collision potential for the highway according to the traffic state, and can be widely used for roadway traffic safety improvement.