Advanced Proactive Road Safety Management Techniques

迫切需要改进道路安全分析方法，以减少道路碰撞的频率和严重程度。目前大多数道路安全研究都集中在统计技术上，以模拟碰撞和评估安全对策。很少有人关注提高我们对车祸原因和道路使用者行为的理解。鉴于缺乏对碰撞因素的复杂相互作用，安全措施如何起作用以及它们如何影响道路使用者行为的完全理解，这是不幸的。这项研究的一个主要贡献将是开发新的方法来推进这种理解。该提案侧重于三个问题：i)评估替代安全措施并验证其与事故的关系；Ii)制定包含道路使用者行为要素的基于回避行动的冲突指标；iii)制定一个校准框架，通过增加模拟冲突与现场冲突之间的相关性，加强在安全分析中微观模拟模型的使用。本研究试图从几个方面为交通安全研究做出贡献。它测试并验证了将交通冲突作为替代安全措施的使用，并确认了它们与碰撞的关系。在贝叶斯框架中结合碰撞记录和冲突数据信息的贝叶斯框架的发展是对安全研究的新贡献。同时，新的基于回避行为的交通冲突指标的发展应该为交通冲突技术提供更好地衡量高混合、无组织交通环境中冲突严重程度的能力。该研究还提出了一种新的定量方法，将各种交通冲突度量（时间接近性和基于行为的）整合为一个单一的安全指数。各种冲突指标被假设为不同的，有时具有独立的性质。最终，重要的道路安全处理决策必须基于对潜在安全水平的单一推断。最后，该研究将通过引入校准过程来最大限度地提高模拟冲突和现场测量冲突之间的相关性，从而推进微观模拟模型在道路安全分析中的应用。针对不同冲突类型、交通环境和位置类型，将确定校正后的模型参数值

There is a pressing need for improved methods of road safety analysis to reduce the frequency and severity of road crashes. Most of current research on road safety focuses on the statistical techniques to model crashes and evaluate safety countermeasures. Less attention has been devoted to improve our understanding of crash causes and road user behavior. This is unfortunate, given the lack of complete understanding of the complex interaction of crash factors, how safety measures work and how they affect road user behavior. A major contribution of this research will be the development of novel methods that advance this understanding. This proposal focuses on three issues: i) evaluating surrogate safety measures and validating their relationship with crashes; ii) developing evasive action-based conflict indicators that incorporate elements of road user behavior and iii) developing a calibration framework for enhancing the use of micro-simulation models in safety analysis by increasing the correlation between simulated and field conflicts. The research attempts to contribute to traffic safety research in several ways. It tests and validates the use of traffic conflicts as surrogate safety measures and confirms their relationship with crashes. The development of a Bayesian framework for combining information from crash records and conflict data in a Bayesian framework is a new contribution to safety research. As well, the development of new evasive-action based traffic conflict indicators should provide the traffic conflict techniques with the ability to better measure conflict severity in high-mix, less-organized traffic environments. The research also proposes a new quantitative methodology for the integration of various measures of traffic conflicts (time proximity and behavior based) into a single safety index. Various conflict indicators are hypothesized to be of different and sometimes of independent nature. Ultimately, important road safety treatment decisions must be taken based on a singular inference on the underlying level of safety. Finally, the research will advance the use of micro-simulation models in road safety analysis by introducing a calibration process to maximize the correlation between simulated and field-measured conflicts. Calibrated model parameter values for various conflict types, traffic environments and location types will be determined.**