公路平曲线段路侧事故预测与风险评价方法研究

发布时间：2019-01-20 13:19

【摘要】：在各等级公路的交通事故中,路侧事故的死亡率极高;我国公路路侧安全形势并不乐观。路侧安全水平的提高需要对路侧事故风险进行准确的评估以便采取针对性的措施,这一研究也可为合理设置路侧净区提供理论基础。本文从公路线形、路面条件等因素入手,基于汽车行驶理论构建了公路平曲线路段汽车不偏离车道驶入路侧的最高安全车速计算模型,通过利用正态分布函数,构建公路的速度分布预测模型。将最高安全车速对应于速度分布累积频率曲线上某一百分位车速,就可以得出交通流中车速高于最高安全车速的比例,即车辆发生驶入路侧事故的概率,从而构建了其与行驶车速、公路线形指标及附着系数的多元关系模型,并进行了实例分析。为了划分影响平曲线段路侧事故发生因素的阈值,选取了路侧事故形态为指标;基于历史的事故数据以及事故模拟再现软件获取的仿真数据,通过设置车辆不同的驶入车速和路侧的环境变量对事故结果数据开展定量的分析,根据车辆驶入路侧后的状态将路侧事故划分成四个严重程度等级:未发生轻微离地、发生轻微离地或坠车、发生至多两次翻车、发生两次以上翻车事故。基于平曲线段路侧事故致因阈值的划分,建立平曲线段路侧事故风险评价的贝叶斯网。通过计算路侧事故致因导致不同形态路侧事故发生的条件概率,代入路侧事故严重程度概率计算的贝叶斯网拓扑结构中,可以得到不同严重程度路侧事故发生的可能性;基于灰色聚类理论对路侧存在障碍物时进行风险评价,并给出了其阈值的划分标准;并通过选取典型车型的路侧事故进行案例分析,将贝叶斯网的推理结果与路侧事故模拟试验结果进行对比,证明了提出的风险评价方法的准确性。论文研究成果预期可为道路交通规划、道路交通工程建设与维护工程提供理论依据与决策参考,有利于改善我国公路路侧安全。
[Abstract]:The death rate of road side accidents is very high in all kinds of highway accidents, and the safety situation of highway side in China is not optimistic. To improve the safety level of road side, it is necessary to accurately evaluate the risk of road side accidents in order to take targeted measures. This study can also provide a theoretical basis for the rational setting of road side clear area. In this paper, based on the theory of vehicle driving, the calculation model of the maximum safe speed of the vehicle on the road side without deviating from the lane is constructed based on the road alignment and road conditions. By using the normal distribution function, this paper presents a new method for calculating the maximum safety speed of the vehicle in the flat curve section of the highway. The prediction model of highway speed distribution is constructed. When the maximum safe speed is corresponded to a percentile speed on the cumulative frequency curve of the velocity distribution, the proportion of the speed in the traffic flow higher than the maximum safe speed can be obtained, that is, the probability of the vehicle coming into the road side accident. The multivariate relationship model between the model and driving speed, highway alignment index and attachment coefficient is constructed, and an example is given. In order to divide the threshold value of the factors affecting the road side accident in the horizontal curve section, the road side accident pattern is selected as the index. Based on the historical accident data and the simulation data obtained by the accident simulation software, the accident result data are quantitatively analyzed by setting different vehicle entry speed and road side environmental variables. According to the state of the vehicle entering the side of the road, the road side accident can be divided into four levels of severity: no slight deviation, slight landing or falling down, up to twice overturning and more than two overturning accidents. Based on the threshold of road side accidents, a Bayesian network is established to evaluate the risk of road side accidents. The probability of road side accidents with different severity can be obtained by calculating the conditional probability of road side accidents caused by road side accidents and substituting the probability of road side accidents into the Bayesian network topology which calculates the probability of road side accidents. Based on the grey clustering theory, the risk evaluation is carried out on the side of the road when there are obstacles, and the criterion of the threshold value is given. The results of Bayesian network reasoning are compared with the results of the road side accident simulation test, and the accuracy of the proposed risk assessment method is proved by selecting the typical vehicle type of road side accident for case analysis. The research results of this paper are expected to provide theoretical basis and decision reference for road traffic planning, construction and maintenance of road traffic engineering, and to improve the safety of highway side in China.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U491.31

【参考文献】