Dynamics of in-station time within metro systems: Measurement and determining factors

Abstract

Worldwide, people living in mega cities are increasingly dependent on metro systems. Their travel experience, however, has not been closely examined. In particular, travel time estimates often do not consider in-station time, which can be significant, especially at large interchange stations with multiple exits and platforms. This study represents a novel attempt to measure in-station time dynamics systematically, considering a wide range of factors such as station design and layout, passenger volume and interaction, and operational schemes. An agent-based modelling approach is used to simulate movement dynamics within metro stations. Then, a robust quantile regression model is built to capture the variability of in-station time and analyze the underlying factors. Four operation scenarios are simulated for the weekday peak, the weekday non-peak, the weekend peak, and a festival holiday peak at two major metro stations in Hong Kong. The findings reveal that the in-station time distribution is the longest during the festival holiday peak, followed by weekday non-peak, weekend peak and then weekday peak. The in-station time varies from 2.5 to 27.5 min, which represents up to 10 times of the in-vehicle time for metro trips within the urban core. Based on the findings, the study recommends both long-term measures, such as increasing the number and density of entrances/exits, and short-term measures, such as providing more escalators at entrances/exits, augmenting the number of inbound ticket gates, improving the experience of transfer passengers, streaming flows to escalators at platforms, and optimizing headways. By adopting these measures, the goal of improving in-station time and travel experience can be achieved more effectively. Overall, this study provides valuable insights into in-station time dynamics and highlights its importance in travel time estimations. This study makes a methodological contribution of developing an agent-based model that takes into account the total passenger experience in relation to the station design and layout, train schedules, operations management and passenger characteristics, such as the total volume, walking speed, trip origins, trip destinations and their interactions.