IEEE Technology and Society Magazine - December 2019 - 68

In the aftermath of large-scale
disasters such as the 2017 Grenfell
Tower fire, understanding human
behavior is becoming recognized
as an important element in the
design process.
Hostile Emergencies
There are similar questions raised for nonnatural emergencies, such as terrorist attacks. These incidents come
in many forms, from active shooters to hostile vehicles
and bombs, each representing a different scenario that
needs to be considered as unique. Each of these incident
types have their own impact on the behavior of people
caught up in them: one might initiate running behavior,
whereas another might induce a state of shock.
Engineers have been approaching the problem of
"hardening" infrastructure from a physical standpoint for
long time (for instance, blast proofing windows, or introducing vehicle bollards), but there has been little work
done on interventions aimed at behavioral change. The
issue remains that, while the benefits provided by the
physical interventions are (relatively) easy to understand,
behavioral interventions are not yet fully understood. In
fact, civilian reactions in terrorist attacks are not very well
understood at all. The fight/flight/freeze/fight paradigm is
easy enough to parrot, but the variables that predict one
reaction or the other have not been fully investigated.
Pedestrian dynamics approaches provide the opportunity
to model the impact of these interventions before an
attack occurs, but require data to calibrate the models.
Without this data the models will remain invalidated, adding little to the "rule-of-thumb" approach.
Obtaining data to understand how people react in unexpected, hostile emergencies is inherently difficult. Before
modern ethics boards, experiments that investigated
responses to stressful situations were fairly common [27],
[28], and some even approached an experimental design
that would have proven insightful for this type of problem
[29]. However, these experiments were performed before
pedestrian dynamics was fully established as a field of
research, and before the advent of the technology that
would allow the extraction of sufficiently granular data for
the calibration of a pedestrian dynamics model.
Outside of experiments, real-world video footage represents the most useful source of quantitative data available. However, this footage is necessarily sensitive, and

isn't routinely made accessible to researchers. Moreover,
it is usually incomplete data, and lacks the experimental
controls that allow effects to be measured.
If this data were to become available, either through
experimental research or through cooperation with the relevant authorities, pedestrian dynamics could make significant advances in safety design for these types of
emergencies. However, in order to obtain this data, ethical
questions will need to be answered. From an experimental viewpoint, the tradeoff between ethical viability and
ecological validity needs to be resolved. From a design
viewpoint, difficult questions about acceptable risk will
need to be answered, and the design profile for future or
nonstandard demographics will have to be well defined.

Supporting Future Work in
Pedestrian Dynamics
Pedestrian dynamics provides an extension to current
design tools, and if used properly can help significantly in
making infrastructure more usable, safe, and secure. It
represents one of the greatest opportunities to design
Smart Cities, to accommodate individuals rather than
masses, and therefore to provide a tailored experience
that, in the most extreme cases, could save lives. However, with that opportunity comes an ethical risk. If this tool
is used inappropriately, or incompletely, then the industry runs the risk of creating problems for the future that
could otherwise have been filtered out in the design
phase. There are three major suggestions that would
support the future work in pedestrian dynamics:
1) Increase The Number and Availability of Open Source
Datasets
There are already open source datasets, but the
industry requires far more data to fully realize
its potential. The primary aim for the industry
and research wings of pedestrian dynamics
should therefore be to gather data surrounding
as many different scenarios as possible and
provide this data as a resource. This data should
be as high quality as possible, providing spatial
coordinates of pedestrians at a high enough
temporal resolution, as well as demographic
data and the details of any surrounding structures. Ideally video footage would be provided,
as well as any further details that influence the
movement of the people in question. All of this
data should be added to a bank of continuously
updated open-source datasets. This would be a
powerful resource for both researchers and
modelers, both as calibrative data and a potential for new research avenues.
2) Develop a Set of Industry Standards and Practices,
and Model Validation Methods

IEEE TECHNOLOGY AND SOCIETY MAGAZINE

∕

DECEMBER 2019

IEEE Technology and Society Magazine - December 2019

Table of Contents for the Digital Edition of IEEE Technology and Society Magazine - December 2019