Counterflow model for agent-based simulation of crowd dynamics

Agent-based crowd models describe pedestrians as autonomous interacting agents. Current models take into account the physical contact forces occurring in a crowd, but the description of many behavioural actions is still a challenge. This paper presents a model for agents’ behaviour in counterflow situations, where they try to avoid collisions with oncoming agents. In the model, the agents observe the walking directions of the agents in front of them and choose their own actions accordingly. We implement the model to the widely used social force model, which describes the motion of each agent in a Newtonian manner. Nevertheless, the basic idea of the counterflow model can be used with various modelling platforms. We study the effects of the model’s parameters with Monte Carlo simulations and justify our selection of their values. Simulation results are compared with previously published experimental data and the results match well.

► We present a model pedestrian behaviour and movement in multi-directional flows. ► The agents observe the others and try to avoid directions with counterflow. ► The presented approach can be applied to any continuous space simulation model. ► Our simulation results match well with existing experimental data.