基于事故替代指标的高速公路匝道与主线结合部安全性能评价及预测方法研究

The methods to evaluate and forecast the safety performance with vehicle speed standerd deviation (SSD) on key poins as the surrogate measure instead of accident were carried out to find a way out of the availability difficulty of history accident data for the traffic researchers and the atypical application of Traffic Conflicts Technique in present China. Firstly, the classificaiton criteria for ramp-freeway junctions were determined and the types mostly constructed in China were classified. And then, the defination of functional area of ramp-freeway was suggested and the range of it was defined using the driver behaviour theory, mathematical statistics, and filed observation. After the calibration of the FHWA's Traffic Software Integrated System (TSIS), the key factors were seleted from traffic operations, geometric desin and signing, and the key poins were also positioned. Furthermore, the relationships between SSDs and key factors were modeled, of which the validity and significance were tested. Finally, the criteria for the level of safety service (LSS) with SSD were provided with the benefit of fuzzy mathematics and other theories.

项目针对我国道路交通科研人员难以获得完整和详细的事故统计资料且交通冲突技术无法典型应用的现状，提出利用特征点车辆车速标准差作为事故替代指标对高速公路事故多发路段- - 匝道与主线结合部的安全性能进行评价和预测的方法。项目首先确定了我国高速公路匝道与主线结合部的类型及分类依据；在此基础上，提出匝道与主线结合部功能区概念，并应用驾驶员行为理论、数理统计、现场交通流数据测量等手段确定其范围；利用标定后的TSIS仿真软件，在实验控制条件下，研究交通流、几何条件以及标志设置等多参数对车辆车速标准差的影响程度，确定关键影响因素，并通过对比分析得出能够表征安全性变化的特征点位置；根据统计要求选取合适的样本点进行现场数据测量，得出特征点车速标准差的预测模型，并验证其有效性和显著性；最后提出安全服务水平概念，利用模糊数学等数学理论确定不同车速标准差所对应的安全服务水平划分标准。

通过大量文献阅读、卫星图片及实地观测，总结出了我国高速公路出、入口匝道与主线结合部的主要类型及其划分依据。应用实测的车头时距数据，计算出了我国高速公路入口匝道与主线结合部加速车道车辆车头时距的概率分布模型，并且在此基础上推导出高速公路加速车道的理论长度计算模型；结合我国工程的实际情况，推导出高速公路出口与入口匝道与主线结合部功能区范围的计算公式，并计算出功能区的长度。利用仿真软件，鉴别出了入口匝道以及出口匝道与主线结合部功能区内车速标准差特征点，并分析出影响特征点车速离散性的关键因素。通过实测数据，得到了匝道与主线结合部功能区内特征点车速标准差均值的预测模型。利用聚类分析等方法，得到了不同安全服务水平下车速标准差的类别和临界值。