Estimating Pedestrian Exposure for Small Urban and Rural Areas

Abstract

Pedestrian exposure refers to a pedestrian’s contact with vehicular traffic that can create opportunities for collisions. A myriad of metrics exists to estimate pedestrian exposure, but no consensus has been reached on what metrics should be adopted. Hence this study attempted to synthesize previous studies and recommend methods for estimating pedestrian exposure, with a focus on rural and small urban areas. Five general types of exposure metrics emerged: from area-based measures (e.g., zonal walk miles traveled), through more granular metrics at the point or segment level, to advanced metrics that utilize the behavioral attributes of walk trips (e.g., space–time prism and discrete choice). In addition, the study used a finite mixture model to estimate a household-level pedestrian exposure measure for rural and small urban settings with the use of the National Household Travel Survey 2009 data. The model accounted for household characteristics (e.g., income and vehicle ownership), regional factors, and block group level attributes (e.g., population density and land use). The results show that the finite mixture model outperformed the negative binomial and zero-inflated models. The results can be used to infer the number of walk trips at as small as the block group level or be inserted into a four-step travel demand model to create point- or segment-based measures where pedestrian network is defined.