he traditional safety evaluation methods are mostly based on historical accident data, which belong to the macroscopical level and have obvious defects in traffic safety management. This paper established accident Probability Density Function by using Kernel Density Estimation (KDE), proposed Accident Probability Prediction (APP) model based on Empirical Bayes (EB) method for considering the impact of accident location characteristics and historical data. The paper also established the method for road traffic safety micro-evaluation by adopting Traffic Accident Probabilities of Equivalent ten thousand Cars (TAPEC) indicator, and a comparative evaluation was conducted by the proposed method against cumulative frequency curve method. Through analyzing accident data collected from the G301, the results show that the proposed method is more reliable and can access to the transformation priorities of potentially dangerous road sections. So it can provide theoretical basis for checking the dangerous sections and improving accident prevention and response.
[1]
Zheng Heng,et al.
A Roadside Safety Assessment Method Based on Bayesian Network
,
2008
.
[2]
Fang Shou.
A New Identification Method for Accident Prone Location on Highway
,
2001
.
[3]
عيسى، محسن العجمي,et al.
Road Traffic Safety
,
2009
.
[4]
Craig Lyon,et al.
Empirical Bayes Procedure for Ranking Sites for Safety Investigation by Potential for Safety Improvement
,
1999
.
[5]
Shaw-Pin Miaou,et al.
Modeling Traffic Crash-Flow Relationships for Intersections: Dispersion Parameter, Functional Form, and Bayes Versus Empirical Bayes Methods
,
2003
.
[6]
Cheol Oh,et al.
Real-Time Estimation of Accident Likelihood for Safety Enhancement
,
2005
.