Available online at www.sciencedirect.com ScienceDirect Transportation Research Procedia 2 (2014 ) 675 680 The Conference on in Pedestrian and Evacuation Dynamics 2014 (PED2014) Choice and experience of the movement route Lee Verhoeff a,b, * a ProRail, Postbus 2038, 3500 GA Utrecht, The Netherlands b University of Birmingham Abstract The standard assumption in both static and dynamic models is that pedestrians will make logical choices based on the shortest time or distance regarding the route that they follow. This approach ignores the fact that perceived time is not identical to clock time, that humans are not logical beings and that they are also influenced by factors such as congestion, crowding, available height, aesthetic quality and the presence of amenities. These factors may help to explain the discrepancies between predicted and observed actual congestion during large scale possessions at Amsterdam Central Station in 2012. 2014 Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/). 2014 The Authors. Published Elsevier B.V. Peer-review under responsibility of Department of PED2014. of Transport & Planning Faculty of Civil Engineering and Geosciences Delft University of Technology Keywords: pedestrian; crowding; congestion; aesthetic; quality; time; dynamic model; environmental psychology 1. Introduction In 2012 two new passageways were built under Amsterdam Central Station connecting the city side of the station to the riverside. Their construction required large scale track possessions which resulted in the other platforms becoming busier as the same number of passengers needed to be serviced while using fewer platforms. In order to identify potential problems, the passenger flows were modelled using the dynamic pedestrian model VISSIM. As VISSIM in that period could not independently determine the choice of platform exit, this choice was defined by expert opinion, which in turn was largely based on the assumption that passengers will use the nearest exit from the platform and will follow the route with the shortest distance. The defined paths were then used as input for a model of the 2011 situation which then was calibrated on the basis of observations of the actual situation. * Corresponding author. Tel.: +31 6 52179189 E-mail address: lee.verhoeff@prorail.nl 2352-1465 2014 Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/). Peer-review under responsibility of Department of Transport & Planning Faculty of Civil Engineering and Geosciences Delft University of Technology doi:10.1016/j.trpro.2014.09.073
676 Lee Verhoeff / Transportation Research Procedia 2 ( 2014 ) 675 680 The impact of all 7 possession periods on the pedestrian flows were predicted using this model. Most problem locations were accurately predicted by VISSIM with the nuance that the model tended to predict level F directly before platform exits while the reality tended to be level E with the same number of passengers being spread over a larger area. There was one notable exception to the accuracy of predictions: a serious problem that was predicted in the first building phase on the 5th platform at the exit to the eastern tunnel. Instead the usual 6 trains per hour with approximately 1700 arriving passengers, 8 trains per hour with approximately 3600 arriving passengers were planned. VISSIM predicted numerous situations such as shown in Fig. 1 at this exit; strangely enough this problem never materialised. Fig. 1. The situation at the exit from the 5 th platform to the eastern tunnel on October 11, 2010 at 8:39:14 and 8:40:16. The gentleman with the light coloured coat has just made it to the top of the staircase in the second photo. In the period since this initial study, it has become possible to analyse the actual passenger routes on the basis of chipcard information. In Fig. 2 the predicted and actual distribution across the tunnels are presented, in the form of closest, middle and furthest tunnel with respect to the stopping position of the train. 100% Percentage of arriving passengers 80% 60% 40% 20% 0% Closest Middle Furthest Closest Middle Furthest Expert opinion input for 2012 model Actual distribution 2013 (chipcard data) Trains on west side Trains on east side Fig. 2. The distribution of arriving passengers on the 5 th platform over the three tunnels in the 2012/3 situation. A later version of VISSIM in which passenger routes are determined on the basis of actual walking time has been since used to model the expected configuration in 2020 of Amsterdam Central Station. As a new direct connection from the Middle to the metro has been incorporated in this model, the results cannot be directly compared to distribution in Fig. 2. As the distribution for the 2020 situation was initially distributed on the basis of a similar expert opinion approach as was used for 2012, these results have been compared in Fig. 3. This comparison demonstrates that the expert opinion approach approximates the actual walking time in the 2020 situation. For the discussion in this paper, the assumption is made that the distribution by expert opinion also approximates the actual walking time in the 2012/3 situation.
Lee Verhoeff / Transportation Research Procedia 2 ( 2014 ) 675 680 677 Percentage of arrving passengers 100% 80% 60% 40% 20% 0% Western tunnel Middle tunnel Eastern tunnel Expert opinion input for 2020 model Predicted by model 2020 Fig. 3. The distribution of arriving passengers from all platforms over the three tunnels in the 2020 situation. The following observations can then be made on the basis of this data: the percentage of passengers arriving on the east side that make use of the eastern tunnel is significantly less than is predicted on the basis of walking time and expert opinion; the distribution of passengers arriving on the east side is significantly different than that of passengers arriving on the west side; and the actual use of the western tunnel (both as closest and as furthest tunnel) is higher than is predicted by on the basis of walking time and expert opinion. These observations suggest that walking time is not the only factor that influences the route choices made by individual passengers and that the western tunnel is more attractive for passengers than the eastern tunnel. Other factors that may influence the route choices and explain the preference for the western tunnel are: congestion and crowding along the entire walking route; available height in each tunnel; aesthetic quality of each tunnel; available amenities and retail offerings along each route; and available destinations from/via each tunnel. 2. Time and Generalised Cost Function Perceived time is a personal experience, more dependent upon how you interpret the events than anything else (Norman (1992)) and has no straightforward relationship with clock time. Time is perceived by attention to, or apprehension of, change through the integration of a series of stimuli and characterized by the ability to conceive of duration, simultaneity, and succession (Li (2003)). Despite this discrepancy between perceived and clock time, pedestrian models assume that pedestrians make their choices on the basis of clock time. The Generalised Cost Function (GCF) is a measure that is used in transportation economics to provide a measure of the cost in time and money of a journey (Balcombe (2004); Wardman (2008) and Department for Transport (2011)). This is achieved by defining weighting factors that convey the influence of the perceived extrinsic benefits, the perceived intrinsic benefits and the perceived effort of a journey. The GCF translates all of these effects, including the monetary cost of a trip, into a time cost ; in this equation a journey of lower value will have a higher time cost. This suggests that such weighting factors might provide the key to explaining the discrepancies observed at Amsterdam Central Station.
678 Lee Verhoeff / Transportation Research Procedia 2 ( 2014 ) 675 680 3. Factors that may influence the route choices 3.1. Congestion Fruin (1987) wrote that humans value personal space. As they acquire status and wealth they select larger offices, automobiles and homes. The pedestrian has his own personal space preferences which are related to his sense of territory and body image. If freedom of choice exists, pedestrians will adopt personal spacing which avoids contact with others. A study by Douglas Economics (2008; See Fig. 4) demonstrated the negative influence of higher service levels (levels E and F) on perceived time. As congestion increases, so those the perceived time spent walking. 8 6 C D E F 4 2 0 0,0 1,0 2,0 3,0 4,0 Weighting factor Density: pedestrians per square meter 3.2. Crowding Fig. 4. Influence of congestion on perceived time spent walking. While density and crowding describe aspects of the relationship between the number of people and available space, they are not identical concepts. Density is a measure of the number of people occupying a defined area while crowding is the experiential state... in which the restrictive aspects of limited space are perceived by the individuals exposed to them (Stokols (1972)). The experience of crowding is reportedly the most common form of discomfort for the commuter and is viewed by many as being stressful (Cox et al. (2006); Turner et al. (2005)). Individuals exposed to crowding have been rated as being more tense and uncomfortable than persons who experienced low density (Holahan (1986)). The effect observed at Amsterdam Central could be influenced by passengers choosing a longer route in order to avoid experiencing crowding. Fig. 5. (a) western tunnel; (b) eastern tunnel. 3.3. Available height The effect of crowding can be mitigated by environmental and architectural factors. It has been demonstrated in that an increase in height (and thus volume) will make a space feel larger and thus less crowded (Baum and Davis (1976); Schiffenbauer et al. (1977)). As the ceiling in the eastern tunnel is significantly lower than is the case in the
Lee Verhoeff / Transportation Research Procedia 2 ( 2014 ) 675 680 679 other two tunnels, a density similar to one in the other tunnels will likely be experience as being more crowded; this could then increase the tendency of some regular passengers to choose a route that feels less crowded. 3.4. Aesthetic Quality The fact that we all in some degree experience feelings of claustrophobia and agoraphobia indicates the very real way in which architectural space may affect us (Holgate (1992)). Environmental psychology literature suggests that individuals react to places with two general, and opposite, forms of behaviour: approach and avoidance. Approach behaviours include all positive behaviours that might be directed at a particular place, such as desire to stay, explore, work, and affiliate (Bitner (1992)). Although little environmental psychology literature is directly focussed on the station environment, useful insights can be found in studies on the retail and shopping mall environments. Yoo et al. (1998) reported that the pleasure resulting from exposure to a store atmosphere influenced such in-store behaviours as spending levels, amount of time spent in the store, and willingness to visit again. The opposite reaction can also be induced: shoppers felt angry, ignored, and displeased in facilities perceived to be unaccommodating. Similar to behaviour in a retail environment it is possible that the higher aesthetic quality of the western and middle tunnels influences some passengers to follow a longer route to avoid the eastern tunnel. 3.5. Amenities and Retail Offerings An important reason (other than that of revenue generation) to offer amenities and retail offerings in a station is to make the visit more pleasurable, thus reducing the perceived time spent there. Pleasure in this context can be influenced by (Ng (2003); Reimers and Clulow (2004)): sensory stimulation; something to do, something to look at; variety of activities (in shopping malls variety has a strong effect on the desire to stay); fulfilment of needs; and provision of convenience. As the eastern tunnel offers no amenities or retail offerings, the route choice of some passengers could be influenced by the perceived sensory stimulation or their desire to either make a purchase. 3.6. Available Destinations At the time that the studies of passenger flows cited in the first section of this paper were made, the only entrance to the metro was situated near the western tunnel; since May 2014 a second entrance near the eastern tunnel has been opened. This together with the bus station and more departing trains on the west side (the midpoint of the platforms is near the western tunnel) meant that more destinations were directly available from the western tunnel than from the other two tunnels; in the perception of many passengers the western tunnel is the main tunnel of the station. This combination of more destinations with the perception of the western tunnel being the main tunnel, may make the western tunnel attractive to more passengers than those who indeed have a shorter route through this tunnel. Conclusion There is a significant difference between the predicted distribution of passengers arriving at Amsterdam Central Station across the three passenger tunnels and the actual distribution as measured using chipcard data. Distributions made by expert opinion and calculated by the dynamic model VISSIM are based on the assumption that passengers will always choose the shortest route in time or distance. The fact that there is a significant difference between these predictions and the actual measured distribution suggests that clock time might not be on its own the ideal measure
680 Lee Verhoeff / Transportation Research Procedia 2 ( 2014 ) 675 680 for determining route choice. A modification where the difference between clock time and perceived time is accounted for might improve the accuracy of these predictions. Environmental psychology research suggests that the experience of the surrounding environment affects the disposition and behaviour of the persons present. As literature directly focussed on the effect of the built environment on passenger behaviour is limited, it is necessary to draw on research from fields such as the retail environment. For the station environment, further research could provide insights into the factors that influence the delta between clock time and perceived time thus leading to improved accuracy of pedestrian models. Acknowledgements The author wishes to acknowledge the advice and support of Dr. Felix Schmid and Mr. Hiro Aso MA Dip Arch RIBA FRSA FCIHT during the writing of the author s dissertation Quality of Stations upon which sections 2 and 3 of this paper is based. All photographs are by the author with the exception of Fig. 5b which is by Michel Kievits. References Balcombe, R., ed. 2004. The demand for public transport: a practical guide. TRL Limited. Baum, A., Davis, G. E., 1976. Spatial And Social Aspects Of Crowding Perception. Environment and Behavior 8 (4), 527-544. Bitner, M. J., 1992. Servicescapes: The Impact of Physical Surroundings on Customers and Employees. The Journal of Marketing 56 (2), 57-71. Cox, T., Houdmont, J., Griffiths, A., 2006. Rail passenger crowding, stress, health and safety in Britain. Transportation Research Part A 40, 244-258, p.249. Department for Transport, 2011. Variable Demand Modelling Scope of the Model Transport Analysis Guidance Unit 3.10.2. Department for Transport, London. p.23. Douglas Economics, 2008. Value and Demand Effect of Rail Service Attributes; Report to RailCorp (Revised Edition). Wellington. p.77. Fruin, J.J., 1987. Pedestrian Planning and Design (Revised Edition). Mobile: Elevator World, Inc., p.20. Holahan, C.J.,1986. Environmental Psychology. Ann. Rev. Psychol. 37: 381-407, p.393. Holgate, A., 1992. Aesthetics Of Built Form. Oxford: Oxford University Press, p.90. Li, Y-W., 2003. Evaluating the Urban Commute Experience: A Time Perception Approach. Journal of Public Transportation 6 (4), 41-67. Ng, C. F., 2003. Satisfying shoppers psychological needs: From public market to cyber-mall. Journal of Environmental Psychology 23, 439-455, p.446. Norman, D.A.,1992. Turn Signals Are The Facial Expressions Of Automobiles, Reading: Addison-Wesley Publishing Company, Inc., p.100. Reimers, V., Clulow, V. (2004) Retail concentration: a comparison of spatial convenience in shopping strips and shopping centres. Journal of Retailing and Consumer Services 11, 207-221. Schiffenbauer, A.I., Brown, J. T., Perry, P. L. et al. (1977) The Relationship Between Density And Crowding Some Architectural modifiers. Environment and Behavior 9 (1), 3-14. Stokols, D., 1972. On The Distinction Between Density And Crowding: Some Implications For Future Research. Psychological Review 79 (3), 275-277. Turner, S., Corbett, E., O Hara, R., White, J., 2005. Health and safety effects of crowding. Rail Safety & Standards Board, Harpur Hill. 14-17. Wardman, M., 2008. Is Generalised Cost Justified? Institute for Transport Studies, University of Leeds, Leeds. pp. 3-4 Yoo, C., Park, J. and MacInnis, D.J., 1998. Effects of Store Characteristics and In-Store Emotional Experiences on Store Attitude. Journal of Business Research 42, 253-263.