Using Cellular Automata to Model High Density Pedestrian Dynamics

The article presents a model of pedestrian dynamics based on Cellular Automata, dedicated especially for high density conditions. Using the proposed representation, it is possible to map different behaviors of agents in various conditions. The article describes mechanisms of a pedestrian's movement for different pedestrian's shapes when a pedestrian occupies multiple cells, compared to the classical representation where a pedestrian occupies a single square cell. A greater number of cells per pedestrian can describe a pedestrian with e.g. a backpack, a suitcase on wheels, or on a bicycle more precisely than when s/he is represented by one cell. More accurate discretizations are particularly useful when high density of pedestrians occurs. The article focuses on the following aspects of a pedestrian's movement: changing direction, movement speed, a passing scheme, and an overtaking scheme.