Macro Level Safety Modeling and Impact Factor Analysis

Abstract:

Due to the lack of systematic research for traffic safety at the planning stage, the relationship between macro level impact factors such as road network patterns, traffic features and other variables with transportation safety remains unclear, which leads to the shortage of available methods and tools to evaluate the safety of various transportation plans. Based on the data from the Orange County, Florida, this paper describes the system research on macro transportation safety modeling at traffic analysis zone (TAZ) level, abstracting effective indices of road network patterns, traffic features and other macro level impact factors. Bayes Space Static Model was adopted to investigate the association between transportation safety and impact factors. Adjacent TAZs are always similar in economic development, road network structures, and traffic features; while TAZs located separately tend to be more diversity. Bayesian Conditional Autoregressive Regression (CAR) Models were used to analyze these spatial correlated data. Considering the different crash mechanisms of major roads and minor roads, new modeling strategy was proposed by modeling crashes on arterials and on local roads separately. The paper finds out there is a significant relationship between the road network patterns and safety, and the new modeling strategy is able to differentiate the safety effect for different types of crashes.