Automatic pedestrian network generation

Abstract

To model pedestrians’ mobility in GIS, the traditional approach used to model transport networks based upon streets centre-lines must be improved. Nowadays, a new level of network detail is needed, representing accurately the real objects used for pedestrians (and their attributes) instead of the ones used for vehicles. This leads to a world of possibilities and applications barely explored. This paper analyses the existing need for pedestrian networks and the scarcity of its availability. The effects of using street networks for studying and/or analyzing pedestrians’ mobility, as it is commonly done to date, are also discussed. To solve this problem, a method to automatically derive pedestrian networks based upon existing geo-data is proposed. The inputs (streets and building blocks) and outputs of the method are presented and critically analyzed.