Human Perceptions of Exit Signs during Building Evacuations
By Robert C. Till, Ph.D., P.E. and Dominick R DeRubbio
Pedestrian simulation and crowd modeling techniques are widely used in the planning and design of structures to increase their level of safety as well as evaluate the level of safety in an existing structure. These models were developed using research regarding pedestrian speeds, the relationship between pedestrian flow and density, and critical density thresholds. While these models are useful for examining pedestrian spaces, they generally do not take into account environmental factors that could influence pedestrian movement such as signage and general route choice decisions. For example, they do not always predict 1) how pedestrians actually behave when searching for the nearest exit, 2) when they take cues from each other or 3) if they follow signs or use some other unknown navigation method.
Fruin used external and overhead movie cameras to record and extract pedestrian flows in subways . More recently, others have used various sensor technologies to model the flow of individuals [3, 4]. In addition, eye tracking has been used to determine how pedestrians avoid collisions [5, 6]. Despite these efforts, little is known about what observers actually look at during the exit or evacuation process.
Despite the presence of exit signs, it is a generally accepted principle in that in an emergency most people will first attempt to leave a building the same way they entered [7, 8]. However, this type of behavior may not be possible in situations such as:
- Mass transit stations where passengers enter a car at one station but due to an emergency are required to exit at a different station.
- High-rise buildings where elevators are used as the primary means of occupant entry to a given floor, but where stairs are used in an emergency.
- Any situation where a person gained entry to a facility via one entrance, but is forced to find an alternative route to exit.
In 2004, Fahy  indicated “exit choice decisions, which determine travel paths and affect travel times” are important data needed for evacuation models which are currently lacking. Fahy also stated that recorded “observations of actual evacuations are ideal, since they show exactly what people did, and the elapsed time can be directly calculated.”
Capturing a video record from the perspective of the observer provides an “ideal” output format. Eye-tracking software goes a step further, supplying a marker on the video scene indicating where the subject is looking.
Using a vision-tracking device, we can get a better idea how people make decisions as they navigate through complex environments. The device tracks the subjects’ eye movements through one camera and the scene through another. Both cameras are mounted on a wearable headset. Software is then used to link the two collected video streams so crosshairs appear on the output video. The resulting output video can show exactly where test subjects are looking as they move through an unfamiliar space.
A pilot study performed in 2011 (10 subjects tested under John Jay IRB protocol JJ-09-152) provided anecdotal evidence that people who have experience moving through a particular environment do not use any signage at all . This group was referred to as “expert” [Video Link]. This was in contrast to “novices” who had no experience moving through the structure, some of whom did make use of signage while navigating.
While further research with 40 more subjects in 2014  did not conclusively confirm the novice/expert dichotomy, both the pilot and subsequent studies did demonstrate that some people do, in fact, make use of exit signage.
Recorded eye movements can depict a subject’s tracking of other pedestrians, architectural treatments, signage and environmental graphics, waypoints and other landmarks and decision points. Analysis of this data can help to determine which environmental cues pedestrians find most beneficial in terms of most efficient use of space and ease of egress. It can support infrastructure planning and design, wayfinding and signage, environmental graphics, simulation of pedestrian movements, as well as evacuation and emergency egress signage and scenario testing. In the future, we hope to be able to draw inferences and conclusions about the fundamental set of behaviors that occur as pedestrians move through a space. This research can help us to develop safer buildings and infrastructure.
Robert C. Till is with the John Jay College of Criminal Justice
Dominick R DeRubbio is with the Agile Risk Group
1Fruin, J. J., Pedestrian Planning and Design, Revised Edition, (Alabama: Elevator World, Inc., 1987).
2Fahy, R., “Available Data and Input into Models,” (paper presented at the Workshop on Building Occupant Movement During Fire Emergencies, Session 4.3, Gaithersburg, MD, 2004).
3Liu, X. Song, W. and Zhang, J., “Extraction and Quantitative Analysis of Microscopic Evacuation Characteristics Based on Digital Image Processing,” Physca A, 388, 2717-2726 (2009).
4Boltes, M. Seyfried, A. Steffen, B. and Schadschneider A., “Automatic Extraction of Pedestrian Trajectories From Video Recordings.” In: Pedestrian and Evacuation Dynamics 2008, Springer-Verlag, Berlin (2008).
5Fujiyama, T., Investigating Density Effects on the “Awareness” Area of Pedestrians Using an Eye Tracker, CTS Working Paper (2006/4), Centre for Transport Studies, University College, London (2006).
6Jovancevic, J. Sullivan, B. and Hayhoe, M., Control of Attention and Gaze in Complex Environments, Journal of Vision, 6, 1431-1450 (2006)
7McClintock, T. Shields, T.J, Rutland, T.R., and Leslie, J., Dishabituation and Stimulus Equivalence Could Make All Emergency Exits Familiar, Journal of Applied Fire Science, 9(2), 125-134 (2000)
8Sime, J.D., Crowd Psychology and Engineering, Safety Science, 21, 1-14 (1995)
9Till, Robert, and Jason Babcock., “Proof of Concept: Use of Eye-Tracking to Record How People Use Exit Signage.” In: Pedestrian and Evacuation Dynamics, Springer, New York (2010)
10Derubbio, D., “Human Perception of Exit Signs During Building Evacuations.” Masters Thesis Submitted to the Department of Security, Fire and Emergency Management, John Jay College (2014).