Evaluation of Pedestrian Density Distribution with Respect to the Velocity Response

There are many approaches to evaluate density within pedestrian scenarios including point approximation, Voronoi cells or more sophisticated methods. In this project we focus on the individual density, where each pedestrian is considered as a source of density distribution. A cone can be used as a reasonable shape with its diameter as a blur parameter. Naturally, pedestrians adapt their velocity and path selection with respect to the conditions around them in given range. The correlation of density and velocity was evaluated on laboratory experiment data for all acceptable blur–range combinations. Because of the fact that negative correlation corresponds to a more significant response of velocity to the density, the correlations seem to be a perfect tool to estimate density parameters. Surprisingly, the expected negative correlation was observed only for one segment of pedestrian’s trajectory, observations were much more complex.