Real-time Proactive Road Safety Management Techniques

多年来，道路安全行业一直依赖警方报告的撞车事故来评估撞车风险，实施改善安全的措施，并评估其有效性。虽然已经开发了几种优秀的方法来最好地利用碰撞数据，但鉴于当今技术的巨大进步和智能城市的出现，它们的局限性不再是可以接受的。由于碰撞数据质量和基于碰撞的安全分析方法的局限性，正在使用替代安全措施，如存在“严重”交通冲突或未遂事故。交通冲突技术作为道路安全分析的主动替代工具已经获得了相当大的普及。冲突发生的频率远远高于碰撞，更容易获得，并在安全分析中提供显着的优势。此外，利用冲突进行道路安全分析具有及时性的优点，并避免了需要首先观察碰撞然后预防碰撞的道德困境。该提案的重点是通过开发一种新的评价方法和应用程序，积极主动地利用交通冲突动态评价道路安全，从根本上改变加拿大和世界各地的道路安全管理做法，以应对当前被动式道路安全分析的挑战。这个发现补助金研究计划的一个主要贡献是实时交通冲突为基础的模型，提供即时和连续的安全措施，可用于动态改变交通环境和优化安全的方法，从根本上背离了过去50年的反应式安全管理的发展。该研究还介绍了实时安全模型的三种新应用，包括：a）对易发生碰撞的位置的动态识别，这可以通过改变位置或网络级的交通控制来改进，B）动态识别到达目的地的最安全路线，这可以为整个网络提供显著的安全效益，以及c）开发自适应交通控制算法，通过最小化碰撞风险来优化真实的交通安全。这些应用程序依赖于车辆轨迹作为主要输入，并可以与来自互联和自动驾驶车辆的数据一起使用，这些数据变得越来越可用。模型和应用将导致新的见解碰撞因果关系作为一个功能的交通网络微观操作。这种方法将在全球范围内彻底改变道路安全管理做法。本研究的第二个贡献是深入了解道路使用者防撞机制的仿真框架的发展。该框架应提高我们实施和评估新的安全措施的能力，并将安全专业人员的重点从解决现有的安全问题转移到规划无问题的道路上。

For many years, the road safety profession has relied on police-reported crashes to evaluate crash risk, implement measures to improve safety, and evaluate their effectiveness. While several excellent methodologies have been developed to make the best use of crash data, their limitations are no longer acceptable in light of the massive technological advances available today and the advent of smart cities. Limitations with the quality of crash data and crash-based safety analysis methods are giving rise to the use of alternate measures of safety such as the presence of `critical' traffic conflicts, or near misses. The traffic conflict technique has gained considerable popularity as a proactive surrogate tool for road safety analysis. Conflicts occur much more frequently than crashes, are more readily available and provide significant advantage in safety analysis. Additionally, using conflicts for road safety analysis offers the advantage of timeliness and avoids the ethical dilemma of the need to first observe crashes to then prevent them. This proposal focuses on addressing the challenges of current reactive road safety analysis by developing a new evaluation methodology and applications that proactively and dynamically evaluate road safety using traffic conflicts to fundamentally change road safety management practices in Canada and worldwide. A major contribution of this Discovery Grant research program is the development of real-time traffic conflict-based models that provide instantaneous and continuous safety measures that can be used to dynamically change the traffic environment and optimize safety-a methodology that departs fundamentally from the past 50 years of reactive safety management. The research also introduces three novel applications of the real-time safety models which include: a) dynamic identification of crash prone locations which can be improved by making changes to traffic control at the locations or at the network level, b) dynamic identification of the safest routes to destinations which can provide significant safety benefits for the whole network and c) the development of an adaptive traffic control algorithm to optimize traffic safety in real time by minimizing crash risk. These applications rely on vehicle trajectories as the main input and can be used with data from connected and autonomous vehicles which is becoming increasingly available. The models and applications will lead to new insights into crash causation as a function of transport network microscopic operations. Such a methodology will revolutionize road safety management practice globally. A second contribution of this research is the development of a simulation framework for in-depth understanding of road user collision avoidance mechanisms. The framework should improve our ability to implement and evaluate novel safety treatments and shift safety professionals' focus from fixing existing safety problems to planning roads that will be problem-free.