Real-time Proactive Road Safety Management Techniques

点击翻译按钮获取中文摘要

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.