ABSTRACT
Accurate measurement of pedestrian trajectories is useful in a variety of situations. In retail environments, infonnation about customer behaviour and preferences can be derived from knowledge of customer's movements. In crowded locations (such as sports venues, railway station concourses and so on) pedcstrian movement influences safety concerns. Our primary concern is in the general planning of pedestrian movemcnts within the urban environment (shopping malls, shopping streets, pedestrian concourses and the like). A related project at Napier University (Pedflow1) has been concerned with predicting pedestrian flows in such scttings. Validation and verification of such models requires data of real pedestrian movements, both in the initial formulation of bchavioural rulcs, and in the testing of the resultant modcls. Traditional methods of gathering such data are laborious, and difficult to automate. The overall aim of this project is to automate th is process, allowing for the gathering of large amounts of data for use in model derivation and validation.
