Simulating the Streets of Tomorrow: An Innovative Approach to Shared Space
Nominated for The Neil Mansfield Award
Stuart Gibb, CH2M HILL
An innovative approach to the problem of how to model shared space aimed at inspiring debate and accelerating wider software development.
The demand on our cities is accelerating at a staggering pace.
As we start to envisage how the ‘smart cities’ of tomorrow will look and perform, the challenge of urban mobility is likely to be high on the list of problems to overcome. As engineers, designers and transport modellers, we have a responsibility to develop integrated and innovative solutions capable of delivering sustainable urban growth and wide reaching socio-economic benefits. One such approach is ‘shared space’.
The philosophy of shared space seeks to readdress the hierarchy of road usership by returning a higher level of importance to the pedestrian. This is generally achieved by forcing drivers to interact with their surroundings at a more ‘human’ level, thus creating an environment where pedestrians are considered to be ‘stakeholders’ rather than just ‘users’.
While some would argue that true shared space cannot be modelled, consultants have seen requests for modelling assistance increase significantly in recent years. Despite this upsurge in interest, the modelling community has been slow to react with little or no best-practice guidance currently in existence.
To the untrained eye, the concept of pedestrians and road users occupying the same space seems relatively simple. Trying to replicate and control this behaviour in a model, however, is extremely difficult. While a number of packages are capable of simulating interactions at standard pedestrian crossings, none were designed with shared space in mind. One such tool is PTV Vissim and Viswalk; one of the industry standard suites of programmes which have been used extensively by consultants and clients for well over a decade.
Despite its many strengths, the software does have one significant weakness in that ‘area-based’ pedestrians cannot see ‘link-based’ vehicles and vice versa. While this lack of functionality is not overly important at standard pedestrian crossings, it is extremely relevant when trying to model something like shared space where pedestrians are expected to navigate around vehicles they encounter en route.
This begs the questions of what can area-based pedestrians actually see and how best could these objects be used to ‘trick’ the software into seeing vehicles? The answer to this is simple: by replacing them with groups of other closely-packed area-based pedestrians.
By introducing ‘dummy’ pedestrians to a predefined sub grid of detectors and ‘micro’ areas, it is possible to repel ‘real’ pedestrians into walking around vehicles using the software's existing Social Force Model.
The concept can be best seen in the following animation:
On top of providing stakeholders with a mechanism for visualising potential schemes, the method would also appear to be capable of providing designers with a legitimate means of estimating the ‘throughput’ of shared space areas. This is a first on both fronts.
Not only does the workaround demonstrate ‘scalable’ innovation, but it is expected that it will go on to inspire further research into shared space and help influence and accelerate wider software development in the mid to long-term.
Association for European Transport