Traffic crashes at toll plazas in Hong Kong

Abstract

Poisson regression was used to identify the significant contributory factors to traffic crashes at toll plaza areas in Hong Kong. Information on the crash incidences and traffic volume at the toll plaza areas of ten tolled roads in Hong Kong during 1998-2003 were obtained from the traffic accident database system and annual traffic census of the Transport Department of the Government of Hong Kong Special Administrative Region. These data, together with the geometric and operational characteristics, including toll plaza width, carriageway width, degree of road curvature, road gradient, and toll booth configuration, were incorporated into two aggregated crash predictive models for different traffic directions. The results revealed that the crash likelihood of inbound traffic was increased significantly with downward slope and the crash likelihood of outbound traffic increased with the degree of road curvature.