高峰期电动二轮车驾驶人事故伤害致因及随机参数分析  

Analysis of electric bicyclist injury severity during traffic peak period with random parameters

作　　者：朱彤[1] 朱秭硕 张杰[2] 肖丹蕾 李青[1] ZHU Tong;ZHU Zi-shuo;ZHANG Jie;XIAO Dan-lei;LI Qing(School of Transportation Engineer,Chang'an University,Xi'an 710064,China;Centre of Road Traffic Safety Research,Research Institute of Highway Ministry of Transport,Beijing 100088,China)

机构地区：[1]长安大学运输工程学院,西安710064 [2]交通运输部公路科学研究院道路交通安全研究中心,北京100088

出　　处：《安全与环境学报》2022年第1期271-280,共10页Journal of Safety and Environment

基　　金：国家重点研发计划项目(2019YFE0108000)。

摘　　要：电动二轮车是我国高峰时段最主要的通勤工具之一。为了研究高峰时段电动二轮车驾驶人事故伤害独特的伤害致因机理,首先,通过独立建模检验(Model separation test)发现高峰时段交通事故伤害严重程度数据特征相比平峰时段有明显差异,应针对高峰时段独立建模;其次,基于发生在某市的2142起电动车与机动车相撞交通事故数据,以事故弱势方电动车驾驶人伤害严重程度为因变量,从双方驾驶人属性、车辆属性与碰撞前运动方向及道路环境因素等方面选取25个自变量进行分析,同时采用随机参数模型进行研究,以更好地解释数据中未能观测到的异质性;最后,根据边际效应获取各显著因素对于电动二轮车驾驶人伤害程度的直接影响。结果表明,在高峰时段,电动二轮车驾驶人性别、涉事机动车车型、碰撞前机动车运动方向及道路环境均会对电动二轮车驾驶人伤害程度造成显著性影响,其中,低能见度是服从正态分布的随机参数,即低能见度对电动二轮车驾驶人的伤害严重程度存在异质性影响,此外,机动车转向运动情况下驾驶人受重伤的可能性较低。研究结果为高峰时段二轮车的交通管理对策制定提供了更具针对性的参考。This paper proposes a random parameter Logit model to analyze the related factors of electric bicycle injury severity during the peak traffic period.To ascertain and clarify the main influential factors of the electric bicycle crash injury severity and random parameter distribution,a descriptive statistical analysis is used to analyze the features of the driver,the accident-involved vehicle,and the circumstance condition based on the 2142 actual cases.According to the model separation test,this study demonstrates that there is a statistical significance between characteristics of crash injury severity of peak traffic periods and off-peak periods,suggesting that traffic peak period crashes should be modeled separately.Next,based on the actual cases,electric bicyclist injury severity was studied by selecting 25 variables from the driver,vehicle,movement of pre-crash,and road circumstance characteristics as independent variables.Then this study categorized the accident severities into four levels:no injury,minor injury,severe injury,and fatal,as the dependent variables.Importantly,the model is utilized to investigate the heterogeneity of variables by relaxing the limitation of the assumption of“IIA”.Thus,this study obtained the discussion of random parameter distribution and relevant discussions.Finally,to quantify the effects of significant factors on the electric bicyclist injury severities,this study investigates the marginal effects.The results manifest that electric bicyclist gender,automobile type,pre-crash automobile movement,and different road environment maintain significant effects on electric bicyclist injury severity during traffic peak periods.Moreover,low visibility is a random parameter obeying normal distribution,highlighting that this indicator's heterogeneous impact on driver injury.Moreover,turning movements of automobiles contribute to reducing the possibility of electric bicyclists suffering severe injury.The results of this study provide the theoretical basis for traffic management strategi

关 键 词：安全社会工程 电动二轮车 交通事故 伤害严重程度 随机参数模型 

分 类 号：X951[环境科学与工程—安全科学]