A Role-dependent Data-driven Approach for High Density Crowd Behavior Modeling

Abstract

In this paper, we propose a role-dependent data-driven modeling approach to simulate pedestrians' motion in high density scenes. It is commonly observed that pedestrians behave quite differently when walking in dense crowd. Some people explore routes towards their destinations. Meanwhile, some people deliberately follow others, leading to lane formation. Based on these observations, two roles are included in the proposed model: leader and follower. The motion behaviors of leader and follower are modeled separately. Leaders' behaviors are learned from real crowd motion data using state-action pairs while followers' behaviors are calculated based on specific targets that are obtained dynamically during the simulation. The proposed role-dependent data-driven model is trained on crowd video data in one dataset and is then applied to two other different datasets to test its generality and effectiveness. The simulation results demonstrate that the proposed role-dependent data-driven model is capable of simulating crowd behaviors in crowded scenes realistically and reproducing collective crowd behaviors such as lane formation.