An Integrated Pedestrian and Vehicle Interface Model for Efficient Planning of Shared Space
Nominated for The Planning for Sustainable Land Use and Transport Award
Samya Ghosh, AECOM, Rekel Ahmed, AECOM
A new approach for developing an integrated tool to forecast strategic pedestrian flows within an urban setting, taking into account the street network, surrounding land uses, and interface with highway micro-simulation software.
There is a strong body of research and commercial software available for simulating pedestrian movements when the overall flows and routes are known (e.g. LEGION, VISWALK, MassMotion). However, there is less research currently available on developing forecasts and routeing of pedestrian flows, on wider network, which serve as the inputs to the microsimulation models.
In this paper, we start by reviewing the existing literature on the topic, including established approaches, such as Space Syntax, but also new tools currently being developed to take advantage of large datasets (e.g. cell phone data) to forecast pedestrian flows in urban settings.
Taking account of the previous research we are developing a new tool that will predict pedestrian flows at the neighbourhood scale for different time periods and days of the week. We will illustrate how this tool was implemented as part of an integrated traffic and pedestrian study in a London Borough.
The tool is designed to be flexible and work with varying levels of data quality and availability. The main variables used in the model are based on data that is freely available via Open Street Map: the characteristics of the street network (overall connectivity, block sizes, shortest paths through the network), and major trip attractors (public transport stations, different buildings and land use types). In particular, we will explore how linking of the tool to area wide traffic models can provide more information to the pedestrian/vehicle interaction. We will extract outputs from VISSIM on traffic volumes and speeds for different time periods and days of the week and show how they can impact pedestrian route choices.
We calibrate our model using pedestrian counts at different locations throughout the site, and analyse its accuracy and potential areas of use. In addition to helping city officials make data driven decisions for planning pedestrian and public space interventions, the ability to predict footfall can be valuable in evaluating the relative value and attractiveness of different land use options. The tool would be applicable for any scale and type of development as well as for locations where pedestrian flows and routes are not well documented or large data collection is not feasible.
We conclude by evaluating our model’s predictive accuracy, and highlighting areas of further research especially greater integration with environmental and socio-economic indicators.
Association for European Transport