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。目前关于道路安全的研究大多集中在用统计技术对事故进行建模和评价安全对策。在提高我们对撞车原因和道路使用者行为的理解方面投入的注意力较少。这是不幸的，因为缺乏对碰撞因素的复杂相互作用、安全措施如何发挥作用以及它们如何影响道路使用者行为的全面了解。这项研究的一个主要贡献将是开发新的方法来促进这一理解。这项建议集中于三个问题：i)评估替代安全措施，并验证它们与撞车的关系；ii)开发包含道路使用者行为要素的基于回避行动的冲突指标；iii)开发一个校准框架，通过增加模拟冲突和现场冲突之间的相关性，加强微观模拟模型在安全分析中的使用。这项研究试图从几个方面为交通安全研究做出贡献。它测试和验证了交通冲突作为替代安全措施的使用，并确认了它们与撞车的关系。贝叶斯框架的发展是对安全研究的一项新贡献，该框架用于在贝叶斯框架中结合来自坠机记录和冲突数据的信息。此外，新的基于规避行为的交通冲突指标的开发应该为交通冲突技术提供在高度混合、组织较少的交通环境中更好地测量冲突严重程度的能力。该研究还提出了一种新的量化方法，用于将交通冲突的各种度量(基于时间接近和基于行为)整合为一个单一的安全指数。假设各种冲突指标具有不同的性质，有时是独立的。归根结底，重要的道路安全处理决定必须基于对潜在安全水平的单一推断。最后，这项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，从而推动微观模拟模型在道路安全分析中的使用。将确定各种冲突类型、交通环境和位置类型的校准模型参数值。**

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.**