高速公路交通事故黑点鉴别及影响因素分析

摘要

Based on the 723 recorded traffic accidents on a freeway segment with 34 km in length, the set accidents frequent sections is obtained using statistical method.Black spot or section that is not af-fected by the starting and ending point is defined.Influencing factors are analyzed using the Bayesian method.Three environment factors and three alignment factors are selected and then the joint probability distribution is inferred.Calculation results indicate that identifying black spots based on the number of ac-cidents is the simplest and the most convenient method.However,the threshold value is different due to different freeway conditions.Among influencing factors of black spots,transition days between weekdays and weekends(Monday and Friday)have larger effects on the accident occurrence probability.Long and steep downgrade is the sufficient condition for accidents on the black spots or sections.Bayesian method could calculate the accident occurrence probability under different influencing factors.The probability of single vehicle accidents caused by small passenger cars is 36% in the rainy and transition days.Within non-single vehicle accidents in weekends,vehicles accidents is the most possible to occur on sunny days and the inferred probability is 38%.%基于某高速公路34 km 路段上723起交通事故记录数据，采用事故数方法统计得到事故多发路段集，并提出不受起讫点影响的交通事故黑点（段）定义。结合贝叶斯理论，对各事故黑点（段）进行致因分析。选取3个道路环境因素和3个道路线形因素，得到多因素影响下事故黑点（段）发生交通事故的联合概率分布。计算结果表明：事故数法是鉴别事故黑点最简便的方法，但其定义阈值宜因不同道路情况而异；黑点事故发生的影响因素中，工休过渡日（周一、周五）对事故发生概率的影响程度较大；连续长下坡是该路段上黑点（段）发生交通事故的充分条件。贝叶斯方法可针对不同的影响因素和事故对象，计算出相应的事故影响因素概率值：在工休过渡日和雨天条件下，小型车单车事故发生在6时至12时的概率为36％；而休息日的非单车事故中，不同类型车事故在晴天条件下发生的概率最大，为38％。