Simulating pedestrians through the inner city: an exploration on the benefits for urban planning

Abstract

This research bridges a gap between agent-based modelling and GNSS-tracking. With the use of both literature and tracking data an ABM model was created that simulated the route choice behaviour of pedestrians (leasurists) on the meso-level. The goal was to create a model to run scenarios with changes in the urban planning or design, to see what effect small changes in the road network have on pedestrian movement patterns. In the model the concept of attraction is used; every crossing in the network the agent chooses the most attractive road. The outcome of the model is mapped and compared to the tracking data. The results are that the running model does not look very realistic, but the output maps of the model do resemble patterns that can be seen in the tracking data. The conclusion is that the model needs to be enhanced in a number of ways: agents should have knowledge on the area, should have destinations, time should be added as a concept and agents may be differentiated in a better way( now it was 'inhabitants, regional visitors and visitors from outside the region').