Generating High-Resolution Pedestrian Trajectories Based on Wi-Fi and Bluetooth Tracking in Urban Outdoor Space: A Preliminary Study

Abstract

While several studies have attempted to understand human mobility using Wi-Fi and Bluetooth tracking, few published studies extract pedestrian trajectories in urban outdoor space with high spatial resolution. We propose a novel framework to generate pedestrian trajectories at the segment level in a commercial district where the main challenge is dealing with the irregularity and sparseness of wireless signals. By covering 10 sensor nodes at each intersection in the study area, we identified the locations of Wi-Fi- and Bluetooth-enabled devices as one of 23 locations, including the midpoints of the segments between intersections. After aggregation and detection of stay points, we constructed 23,528 trajectories for 18,191 Wi-Fi-enabled devices and 518 trajectories for 441 Bluetooth-enabled devices. Based on Wi-Fi trajectories, on average, people were found to stay at 2.02 points for brief periods, 1.76 points for long periods, and travel about three times over 373.26m on average within our study area. This work could help transportation planners and policymakers better understand pedestrian movement trajectories and activity patterns.