Simulating Crowd Movements Using Fine Grid Cellular Automata

Cellular automata models have been extensively used for simulating pedestrian movements. One weakness of this type of models, is the size of the cells which is mostly considered to be of the size of a pedestrian (i.e. 40cm*40cm or near to it). Each pedestrian stands in a single cell and during simulation time steps, transits to neighbouring cells. The size of the cells dictates discrete speed steps for pedestrians and chess like movements. A pedestrian could move one or more cells in a simulation time step or do not move at all. In order to have more smooth movement and different speeds for pedestrians (i.e. from different ages, physical and health status) we propose a finer grid cellular automata. We use this model to build a least effort cellular automata movement algorithm. The algorithm has been used in small scale movement layer of our multi-agent crowd simulation platform in order to allow more realistic simulations of crowd flows. A simple walkway scenario has been used to test the outcome of the model. Results of the simulation are reported in this paper.