高速公路车辆跟驰运行安全可靠度及风险控制研究

Freeway rear end accidnet prevention is the key content in Chinese road safety management. For the operating vehicles on the freeway, alarming intervetion is the only countermeasure to prevent the rear end collision. According to the uncertainty of the driver's behavior,it is hard to decribe the safety status of the following vehicle, which is the premise of alarming intervetion. Applying engineering reliability theory to traffic control and management for the rear end collision prevention, this study setups the driver's behavior module and its index under vehicle following environment condition. Basing on the driver simulation and on field experiments, the probabality function of diviation from the standard behavior is set up to sovle the difficluty in human reliability value, which is the impossiblity of carrying out thousands of experiment to get the relaiblity value. Putting the human reliablity model into the traditional vehicle following model, the vehicle following reliabilty model is therefore built, and simulation is utilized to solve the function. Finally calculating the vehicle following reliabilty value change due to the driver intervening alarm, the risk management decision support model of different traffic control measures is set. Therefore the freeway vehicle following reliabilty and risk mangement theory is intially integrated and applied in real traffic managemnt. This study could not only provides the basis for freeway traffic operating safety management, but also is meaningful to the vehicle active safety techique, other type of traffic accident prevention and driver behavior.

高速公路追尾事故预防是我国当前的交通安全保障技术中的重点内容。对于在途机动车，驾驶人预警干预是追尾事故预防的必然手段。但由于驾驶人行为的不确定性，一般研究手段很难解决"定量描述跟驰状态的安全程度"这一预警干预前提。本项研究将可靠度理论应用于人-车-路系统，构建跟驰环境下驾驶人行为模式指标集，实验采集相关指标，采用偏离标准值的概率函数描述驾驶人可靠度，解决"无条件进行数万次以上试验获取可靠度定值"的人因可靠度定量描述难点问题。将其纳入传统跟车模型，行成高速公路车流运行安全可靠度模型和仿真计算方法。最后，从驾驶人可靠度及车流运行安全可靠度干预效果的角度，提出基于追尾风险规避的交通管控措施决策支持方法。从而完整高速公路跟驰运行可靠度及风险控制理论，并在交通管控中初步应用验证。本研究不仅可为高速公路安全运营控制管理提供依据，且对车辆主动安全技术、其它各类交通事故预防和驾驶人行为等方面都有重要意义。

高速公路追尾事故预防是我国当前的交通安全保障技术中的重点内容。对于在途机动车，驾驶人预警干预是追尾事故预防的必然手段。.本项研究将可靠度理论应用于人-车-路系统，构建跟驰环境下驾驶人行为模式指标集，实验采集相关指标，采用偏离标准值的概率函数描述驾驶人可靠度，解决人因可靠度定量描述难点问题。本实验通过不同渠道招募和选择测试驾驶人共30名。采集其感知、判断、操作过程的行为数据，建立可靠度函数。感知阶段失误服从lnN(0.02，0.161)的对数正态分布。速度估计误差失误服从LnN(-1.498，0.269)的对数正态分布，速度估计误差失误服从LnN(-1.498，0.269)的对数正态分布。制动动作时间失误概率服从LnN(0.116，0.284)的对数正态分布。.行成高速公路车流运行安全可靠度模型和仿真计算方法。建立ANFIS跟驰模型。训练结果 RMSE和CORR分别为0.05和0.97.测试结果RMSE 和CORR分别为 0.06和0.95。.最后，从驾驶人可靠度及车流运行安全可靠度干预效果的角度，提出基于追尾风险规避的交通管控措施决策支持方法。研究发现高速公路车流检测器中车速时空矩阵和占有率时空矩阵的特征向量对追尾风险的影响重大，控制车速均衡和占有率为关键因素。通过交通控制，调整车流特征时空矩阵，可以降低事故风险。.本研究不仅可为高速公路安全运营控制管理提供依据，且对车辆主动安全技术、其它各类交通事故预防和驾驶人行为等方面都有重要意义。.本项研究成果：发表SCI/SSCI论文8篇，EI论文3篇，申请发明专利1项，申请软件著作权1项，获得上海市浦江计划人才项目1项。