论文部分内容阅读
Urban arterials feature large traffic volumes and severe traffic safety conditions.For the purpose of improving traffic safety,the key issue is to identify crash occurrence contributing variables,of which speed arouses attention increasingly.Floating Car Data obtained by taxis equipped with Global Positioning System,were used in the study to capture the features of speed along a segment.Based on a total of 234 one direction road segments from 8 arterials,this study aims at analyzing safety performance considering average segment speed.Regarding data structure as arterial-level and segment-level,Bayesian hierarchical Poisson log-normal model was developed.In addition,Maximum Likelihood Estimation prior was adopted to improve estimates reliability.Results indicated average segment speed was negatively associated with crash occurrence; other variables(e.g.density of signal spacing,segment volume,segment length,number of lanes,number of phases,bus-only lane)were also significantly associated with crashes.