Real-time Proactive Road Safety Management Techniques

多年来，道路安全专业一直依靠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大限度地利用撞车数据，但鉴于当今大规模的技术进步和智能城市的出现，它们的局限性不再是可以接受的。碰撞数据质量和基于碰撞的安全分析方法的局限性导致使用替代的安全措施，例如存在“严重”交通冲突或险些发生的事故。交通冲突技术作为道路安全分析的一种主动替代工具，已经得到了相当大的普及。冲突比撞车发生得更频繁，更容易获得，并在安全分析方面提供了显著的优势。此外，将冲突用于道路安全分析提供了及时性的优势，并避免了需要首先观察碰撞才能防止碰撞的伦理困境。这项建议的重点是通过开发一种新的评估方法和应用程序，利用交通冲突对道路安全进行主动和动态的评估，从根本上改变加拿大和世界各地的道路安全管理做法，以应对当前反应性道路安全分析的挑战。Discovery Grant研究计划的一个主要贡献是开发了基于实时交通冲突的模型，提供了可用于动态改变交通环境和优化安全的即时和连续安全措施-这一方法与过去50年的反应性安全管理根本不同。研究还介绍了实时安全模型的三个新应用，包括：a)动态识别易发生碰撞的位置，可通过改变位置或网络级别的交通控制来改进；b)动态识别到达目的地的最安全路线，可为整个网络提供显著的安全效益；c)开发自适应交通控制算法，通过最小化碰撞风险来实时优化交通安全。这些应用程序依赖于车辆轨迹作为主要输入，并可与联网和自动驾驶车辆的数据一起使用，这些数据正变得越来越可用。这些模型和应用将导致对作为交通网络微观操作函数的撞车原因的新见解。这种方法将彻底改变全球的道路安全管理实践。这项研究的第二个贡献是开发了一个仿真框架，用于深入了解道路用户的防撞机制。该框架应该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没有问题的道路上。

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.