DESIGNING AND PLANNING FOR SAFE PEDESTRIAN PATHS AT RAIL TRANSIT STATIONS SERVING RITI COMMUNITIES

Abstract

In this study, we focus on pedestrian network construction and pedestrian route choice analysis. We developed a GIS based framework for pedestrian network construction, which takes multiple data sources, such as open source networks, satellite imagery, and pedestrian GPS traces. The pedestrian route choice study examines the impact from tradeoffs between environmental and infrastructure attributes, such as ambient noise, tree canopy shade, and surface characteristics (e.g., sidewalk, grass, etc.). We investigate these in a university campus setting, where walking trips comprise about 25% of all commute trips, with a greater percentage expected for within campus OD trips. We collect and analyze GPS data from volunteer community members of the University of Hawaii at Manoa (UHM), resulting in 298 distinct observed OD trips and their routes. From a RUM route choice standpoint, choice set generation is a difficult problem, especially for on-campus walking, which is unrestricted and can deviate from discrete roadways or sidewalks. Thus, a recursive logit route choice model is estimated to determine the tradeoffs between route link attributes, such as ambient noise, tree canopy shade, and other infrastructure attributes. The estimated recursive logit model and network construction framework were applied to four identified Skyline stations to analysis the pedestrian route choice behavior when accessing the stations.