Correlation Analysis of Traffic Conflicts and Driving Behaviors at Interchange Diverging Areas

At freeway interchange diverging areas, traffic conflicts are caused directly by driving behaviors. This study aimed to investigate the correlation between traffic conflicts and driving behaviors, which include acceleration/deceleration and lane changing at diverging areas based on Chinese cases. Principal component analysis was used to make a dimension reduction of the parameters that influence traffic conflicts. The Bayesian network was used to construct a correlation model between parameters and traffic conflicts in diverging area. A significant correlation existed between the lane-changing position and the traffic conflicts, which were located within 600 m from the beginning of the diverging area and the first 60 m of the deceleration area. A significant correlation existed between acceleration/deceleration and the traffic conflict. Deceleration and lane-changing behavior are the main factors that affect traffic conflicts. The research provided useful advice for driver training programs to reduce casualties and a basis for the active prevention of accidents.