Empirical Evidence on Platform Passenger Distribution and Dwell-Time Impacts in Urban Rail Networks

This study empirically examines the spatial distribution of passengers on urban rail platforms and its impact on boarding and alighting. It combines two complementary datasets from Vienna: (i) automated passenger counts, which record approximately 4.4 million boarding and alighting movements at 20 stations on the Vienna metro network over the course of one month, and (ii) detailed visual observations at 10 S-Bahn stations, covering approximately 15,000 positions of waiting passengers. The results show that passengers are not distributed evenly along the platform. Regular passengers, in particular, choose which door to use when boarding, based on the shortest expected route when alighting. Instead, stable local clusters form at individual doors, where up to 3.5 times the theoretical average passenger share can be observed. It is noted that this uneven distribution decreases during periods of high occupancy and increases during periods of low occupancy. These distributions prolong dwell time, reduce timetable robustness and refute the widespread assumption of uniform passenger distribution in planning and simulation. This work establishes an empirical basis for door-specific modelling of stopping times, the targeted placement of platform access points and information-based demand management.