Empirical Study on Congested Subway Transfer Traffic Patterns

Empirical Study on Congested Subway Transfer Traffic Patterns

Meng-Cong Zheng

Department of Industrial Design, National Taipei University of Technology, Taiwan

Available online: 
30 September 2018
| Citation



Taipei Main Station of Taipei Mass Rapid Transit is the busiest transport hub in Taiwan in terms of ridership. Its complex layout and high number of passengers frequently lead to congested transfer traffic patterns. This study examined passengers’ walking trajectories and behaviours and the relationship between crowding and train movement at the transfer concourse on floor B2 of Taipei Main Station to understand the factors of interference and congestion during traffic flow. An improvement plan was subsequently proposed. This study observed that because more passengers situated themselves in the middle cars than the front and rear cars, most boarding and alighting passengers used specific escalators to enter and exit the platform level. In addition, passengers’ walking flow tended to be affected by their personal moving distances, the movement of other passengers and traffic volume. Transfer passengers preferred to use escalators or stairs closer to them, resulting in poor traffic diversion inside the platform. In particular, congestion frequently occurred at the fork near the T junction, where most passenger interferences were recorded. Passengers tended to lean against walls or walk between pillars to mitigate the conflicting flow of movement among them. Other walking trajectory factors included the locations and directions of escalators, stairs and turnstiles. This study used Unity3D software to construct three traffic diversion proposals based on observation records. The proposals were used to simulate and verify improved traffic patterns and mitigate interference. The simulations revealed that moderate changes in the upward and downward directions of escalators could facilitate smoother transfer traffic patterns. Escalators with traversing directions that better adhere to passengers’ traffic patterns may substantially increase passengers’ walking speeds regardless of the direction they are coming from, thereby effectively mitigating congestion at the T junction.


non-participant observation, patterns of behaviour, subway transfer, Unity3D


[1] Alexandra, W., Human movement behaviour in urban spaces: implications for the design and modelling of effective pedestrian environments. Environment and Planning B: Planning and Design, 31, pp. 805–828, 2004. DOI: 10.1068/b3060

[2] Hoogendoorn, S.P., Bovy, P.H. & Daamen, W., Walking infrastructure design assessment by continuous space dynamic assignment modeling. Journal of Advanced Transportation, 38(1), pp. 69–92, 2004. DOI:10.1002/atr.5670380106

[3] Wen, K.-C., A study of passenger’s moving conflict to ameliorate flow by simulationbased evaluation in MRT Taipei Main Station. Master thesis, Chinese Culture University, 2011.

[4] Fan, C.-H., A study on passengers weaving behavior model. Master thesis, TamKang University, 2011.