Non-crash Based Safety Analysis: Innovative Models or Urban Intersections and Highways

点击翻译按钮获取中文摘要

The transportation system is in a continual development to meet every increasing demand for mobility, safety, and increased capacity. This is required along with increasing emphasis on sustainability and resource-efficiency. In the quest to create more reliable models and analysis techniques, researchers are collecting as well as analyzing ever-increasing volumes of data. The proposed research supports the development of two frontiers: [i] creating more reliable models for safety prediction without the reliance on crash data and [ii] creating a more unified framework to analyze crash and non-crash data. The first goal of the proposed research attempts to advance the knowledge on how safety can be predicted without resorting to crash data. This involves the search for clues' in the traffic stream which can reflect safety hazards. More interestingly, if such clues can be obtained from hypothetical or simulated traffic facilities. In building these models, a special focus will be allocated to non-motorized models of travel. The transferability of calibrated models will also be investigated in order to explore the generality and applicability of these models. The second goal of the proposed research attempts to develop a unified framework to model crash and non-crash events for urban intersections as well as freeway acceleration lanes. Modelling of driver, pedestrian, and cyclist choices as they navigate traffic facilities will be attempted. Subsequently, it will be investigated whether such model can be used to predict the incidence of crashes on these facilities.