RUI: Behavior-Based Stochastic Traffic Flow Modeling for Intersection Safety Improvement

这项研究将调查信号交叉口的驾驶员行为，并推导出旨在改善交叉口安全的复杂交通流模型。交叉口安全一直是全国关注的问题，部分原因是缺乏对复杂驾驶行为的理解，特别是驾驶员面临信号相位变化时的决策机制。本研究通过对信号交叉口的驾驶员决策进行全面调查来解决这个问题，并利用这些调查的见解，为这种中断的情况建立了一个随机交通流模型。通过将开发的交通流模型与互联车辆技术（特别是V2X）相结合，开发的交通流模型可用于量化信号交叉口的安全性能，识别新出现的危险情况，并有助于驾驶员辅助和交叉口事故避免技术的发展。此外，这项研究的成果将为学生创造机会，特别是那些来自代表性不足的群体，通过交通工程的教育和研究机会。它还将通过课程开发、教育模块和高中与大学之间以及大学与研究生院之间的途径，为大学生、交通工程师和K-12教师和学生提供机会。1）研究复杂的驾驶行为和驾驶员决策的内在机制;通过分析从视频图像中提取的车辆轨迹数据，在信号交叉口处做出; 2）通过考虑驾驶员决策的随机性，建立随机交通流模型，以描述复杂的驾驶行为并预测信号交叉口处潜在的交通冲突。使得当面对信号相位变化时情况发生变化。该研究将有助于交通流模型的理论发展。这些交通流模型将能够描述复杂的驾驶行为和估计信号交叉口的交通冲突，这两个都是大多数其他交通流模型所缺乏的。此外，该研究有望为提高交叉口安全性做出重要贡献。这项研究将为未来开发动态系统奠定基础，以提醒驾驶员注意新出现的危险，并帮助避免交叉路口事故。

This research will investigate driver behavior at signalized intersections and derive sophisticated traffic flow models aimed at intersection safety improvement. Intersection safety has long been a national concern, partly due to the lack of understanding of complicated driving behaviors, especially decision-making mechanisms present when drivers face signal phase changes. This research tackles this issue by conducting a comprehensive investigation of driver decision-making at signalized intersections and, using insights from these investigations, developing a stochastic traffic flow model for such interrupted flow situations. Through the integration of the developed traffic flow model with connected vehicle technologies (specifically, V2X), the developed traffic flow model can be used to quantify the safety performance of signalized intersections, identify emerging hazardous situations, and contribute to the development of driver-assistance and intersection-accident-avoidance technologies. Furthermore, the outcomes from this research will create opportunities for students, particularly those from underrepresented groups, through educational and research opportunities in transportation engineering. It will also provide opportunities for college students, traffic engineers, and K-12 faculty and students through curriculum development, educational modules, and creation of pathways between high school and college, as well as college and graduate school.The objectives of this research include: 1) investigating complicated driving behaviors and the inner mechanisms of drivers' decision-making at signalized intersections through the analysis of vehicular trajectory data extracted from video images and 2) developing a stochastic traffic flow model to describe complicated driving behaviors and predict potential traffic conflicts at signalized intersections by considering the stochastic nature of drivers decision-making when facing signal phase changes under varying circumstances. This research will contribute to the theoretical development of traffic flow models. These traffic flow models will be able to describe complicated driving behaviors and estimate traffic conflicts at signalized intersections, both of which are absent from most other traffic flow models. Furthermore, this research is expected to contribute significantly to the improvement of intersection safety. This research will build a foundation for the future development of dynamic systems for alerting drivers of emerging hazards and helping to avoid intersection accidents.