Using a Latent-Class Model to Examine Sensitivity of Local Conditions on Stair and Escalator Choice in Toronto Subway Stations

Understanding the rationale behind routing decisions for pedestrians in complex spaces remains a technical challenge for space design and crowd management. The ability to correctly predict crowd usage of passages, open spaces and vertical transport (e.g. stairs, escalators) would allow for better design and operation of such facilities. To date, minimal research has been performed examining vertical transport choice in the context of pedestrian path finding, and existing models have been static in nature. This study aimed to better understand the relative weighting pedestrians give to an initial predisposition versus that based on more dynamic conditions (e.g. queuing) at the entrance of the facility. Two types of discrete choice models were estimated, a standard binomial logit, which assumes a homogeneous population, and a latent-class model, which allows for population segmentation. The latter was formulated to consider the process as a combined decision based on information available before and after reaching the facility. A strong aversion to stair use was found, with an additional dampening effect of increasing opposing flow. Conversely, approaching from the staircase side, having preceding pedestrians take the stairs and higher flows or escalator queues were found to increase the chance of stair use. The latent-class model also elucidated a relatively minor influence of dynamic conditions on overall choice. Significant predictive differences were not found between the models; however, the ability of the latent class model to identify the relative weight of local dynamic conditions allows for insight into the decision process of pedestrians.

• 
• Supplemental Notes:
  • This paper was sponsored by TRB committee AP045 Intermodal Transfer Facilities.
• Corporate Authors:

  Transportation Research Board

  500 Fifth Street, NW
  Washington, DC  United States  20001
• Authors:
  • Srikukenthiran, Siva
  • Shalaby, Amer
  • Habib, Khandker M Nurul
• Conference:
  • Transportation Research Board 93rd Annual Meeting
  • Location: Washington DC
  • Date: 2014-1-12 to 2014-1-16
• Date: 2014

Language

• English

Media Info

• Media Type: Digital/other
• Features: Figures; Photos; References; Tables;
• Pagination: 14p
• Monograph Title: TRB 93rd Annual Meeting Compendium of Papers

Subject/Index Terms

• TRT Terms: Choice models; Crowds; Escalators; Pedestrian movement; Routing; Stairs; Subway stations; Traffic queuing
• Geographic Terms: Toronto (Canada)
• Subject Areas: Pedestrians and Bicyclists; Planning and Forecasting; Public Transportation; I72: Traffic and Transport Planning;

Filing Info

• Accession Number: 01515848
• Record Type: Publication
• Report/Paper Numbers: 14-3287
• Files: PRP, TRIS, TRB, ATRI
• Created Date: Feb 25 2014 9:15AM