The unintended consequences of design on pedestrian experience

March 26, 2019

A key part of building design is to assess the effectiveness of design elements in meeting a stated performance objective – whether they do the thing we want them to do.

In relation to people movement (e.g. pedestrian and evacuation dynamics), such assessments are made regarding the impact of a design on either

  • the safety of the population (e.g. during an evacuation from fire),
  • the security of the population (e.g. structural changes preventing marauding terrorist attacks),
  • and the comfort of the population during routine operation (e.g. enhancing comfort during circulation).

Regulatory requirements and guidance constrain or support this assessment depending on the objective. For instance, does an emergency sign convey the required information, does an emergency exit have sufficient capacity to cope with expected users, can a notification system be heard in the target areas, etc? These are design objectives for these elements to affect the safety domain.

Within any of these three core domains (safety, security and operation), such assessment requires the compilation of several factors to form scenarios of interest. The design can then be tested according to these scenarios which represent conditions that might be expected when eventually in use. These scenarios might be formed from population factors, configuration factors,environmental factors and procedural factors. For instance, some buildings may host different types of event requiring the effectiveness of the building design (or some aspect of it) to be examined across these events; e.g., a conference centre may host an array of different events requiring different internal furniture configurations that will affect pedestrian movement and experience. The population size in other structures may vary significantly overtime requiring analysis to assess different sizes; for instance, a railway station will have tidal flows of passengers based on peak times and train timetables, and require performance to be examined across such flows. All buildings will have visitors with different movement, sensory and cognitive abilities that would need to be represented in any assessment. The practitioner identifies the factors that combine to form the scenarios to be examined to assess the impact of the design in the domain.

A particular design change (ranging from a new building to a seat belt) will then need to be assessed across a range of different scenarios to determine whether it meets stated objectives across expected conditions – be it related to safety, security or operations. It is challenging to credibly demonstrate the effectiveness of a design element (e.g. the additional comfort provided by anew chair, the increased probability of identifying an object given a new search procedure, the reduced evacuation time given a new emergency procedure,etc.), and go on to quantify the effect to enable comparison between different design variants.  

It may be possible fora practitioner to derive the impact of a new design element on a small set of safety or security or operational outcomes. That is, an expert examining the design and deriving its impact on performance directly through analysis or from first principles. For instance, an expert may examine the plans for a new door and, from experience, determine how children and the elderly might interact with it. This requires great expertise and precise questions to be asked to enable outcomes to be logically derived.

It may also be possible to conduct small-scale tests to establish this performance. For instance, to build a mock-up of anew turnstile to see its impact on the throughput of arriving participants or observe it in isolation during use. However, these tests are challenging, costly to conduct and, again, typically examine performance in isolation of many oft he factors that might be present in representative scenarios – especially when completed in a laboratory setting.

These approaches enable the impact of single factors to be derived or observed in detail. It is more challenging to assess the interaction between different factors as their number increases and the scenarios become more complex. For instance, as noted above, it is possible to establish the processing time of a new turnstile design – either analytically or from making direct observation. Although this understanding is key, it does not quantify the interaction between the installation of a set of such turnstiles in context with other facilities before and after the turnstiles, different populations, etc. – other, more complex scenarios. It gets you so far, but does not describe these complex interactions to the extent that predictions can be made of queuing levels produced. Similar challenges exist in the security domain. For instance, security concerns may require a new checkpoint design with associated search and screening procedure to accommodate this design. The same security concerns may also require other structural changes to alleviate threat concerns (e.g. the implementation of hostile vehicle mitigation barriers). However, it will be challenging to derive the combined impact of these physical and procedural changes on people movement.

In such situations, expert understanding and observations should be accompanied by pedestrian modelling to generate insights into the complex interactions of many factors. Of course, there is no magic in these models – they are a composite of existing expertise and observations. However, through their computational power they are able to capture the interaction of the many factors that might be present in a scenario of interest for safety, security or operational scenarios. Depending on their scope and refinement, they might be able to produce results across the scenario space (the range of scenarios that might potentially exist) within a particular domain. For instance, looking at the performance of turnstile designs, with different population loads and types arriving, and with different surrounding spatial designs.

It is actually even more complicated than this. A design element (a turnstile, a door, a seat, a procedure, an alarm) is typically introduced to produce a desired effect on some aspect of performance. A new seat is introduced as it is more comfortable and cost-effective; an entrance is repositioned to address access issues; a new evacuation procedure is introduced to clear the building more effectively. A practitioner is typically assigned to assess the success of a design element within a particular domain – safety, security or operational. However, in reality, design modifications may have unintended consequences relating to safety, security and operational domains. These unintended consequences may not always get picked up in focused observations or even in equivalent expert analysis. These activities typically need to be narrow in scope to reduce cost and complexity and isolate the effects of factors. Pedestrian modelling offers the opportunity (not always taken) to look at the side effects of new designs given the range of scenarios that can be examined formed of numerous factors – a degree of complexity that might be beyond experimental observation and direct analysis. These might encompass security, safety and operational concerns. For instance, establishing the impact that the increased throughput of new turnstiles has on the queuing produced downstream within a structure,determining whether the new location of an entrance (introduced to aid access) influences the effectiveness of the evacuation procedure in place, etc. These outcomes might be established through modelled scenarios deliberately representing expected scenario conditions or by conducting explorative scenarios to stress test performance – to establish unforeseen issues that might require further(expert or experimental) analysis.

Pedestrian models do not solve the problem. They may provide users (with expertise in the fundamentals and in data collection) with another tool to help them quantify the impact of different designs and also demonstrate issues on other domains that might otherwise had not been noticed. They need to be sufficiently sophisticated and wielded in expert hands – but, given so, might highlight problems within and between domains that might otherwise not be found until the design has been implemented.  These tools should help provide additional evidence used in making your organisation ready for safety, security and operational changes that affect the movement of people.