ABSTRACT：Urban arterials form the main structure of street networks. They typically have high traffic volume and high crash frequency. Hotspot identification (HSID) is the first step for traffic safety management process and often utilizes crash prediction models. The classical crash prediction models investigate the relationship between arterial characteristics and traffic safety at micro level, since they treat road segments and intersections as isolated units. This micro-level analysis has limitations when examining urban arterial crashes because: 1) signal spacing is typically short for urban arterials in dense street network, and there are interactions between intersections and road segments that classical models do not accommodate; 2) in practical engineering, a hotspot consists of several adjacent intersections and road segments instead of a single intersection or road segment. Taking these into account, a meso-level unit that combined signalized intersections and their adjacent road segments as a whole was adopted. To investigate the suitable research unit and method for urban arterial HSID, this study identified hazardous micro-level (intersections or road segments) and meso-level units at the same time using crash frequency, empirical Bayesian (EB), potential for safety improvement (PSI), and full Bayesian (FB) methods. Consistency was tested to evaluate the performance of the HSID methods. The results showed that 1) meso-level units performed better than micro-level units regardless of which HSID method was adopted; 2) EB and PSI performed better than the other methods no matter for which research unit; 3) there was a big difference between the identified hazardous micro- and meso-level units.
Jia Li, Xuesong Wang. Hot Spot Identification of Urban Arterials at the Meso Level. Transportation Research Board 98th Annual Meeting, Washington D.C., USA, 2019. 1.13-17.