Simulating bi-directional pedestrian flow in a cellular automaton model considering the body-turning behavior

In the experiments of bi-directional pedestrian flow, it is often observed that pedestrians turn their bodies and change from walking straight to walking sideways, in order to mitigate or avoid the conflicts with opposite walking ones. When these conflicts disappear, pedestrians restore and walk straight again. In the turning states, the forward velocities of pedestrians are not affected. In order to simulate this body-turning behavior, we use a cellular automaton (CA) model named ITP model, which has been proposed before. But the occupied area of one pedestrian is set as 0.4 m∗0.2 m. After the introduction of new rules of turnings and restorations, the pedestrians become more intelligent and flexible during the lane formation process, and some improvements of the fundamental diagram of pedestrian flow can be found. The simulation results of two different scenarios under open boundary conditions are also presented, and compared with the experimental data. It is shown that the new model performs much better than the original model in various tests, which further confirms the validity of the new rules. We think this approach is one useful contribution to the pedestrian flow modeling.