Abstract:For walking pedestrians, when they are blocked by obstacles or other pedestrians, they adjust their speeds and directions to avoid colliding with them, which is called collision avoidance behavior. This behavior is the most complex part of pedestrians' walking processes and its modeling and simulation are the keys to realistic crowd simulation, which serves as the foundation for various applications. However, most existing methods either lack the representation power to accurately model the complex collision behavior or do not model it explicitly, which leads to a poor level of realism of the simulation. To realize realistic crowd simulation, we propose to analyze, understand, and model the collision avoidance behavior in a data-driven way. First, to automatically detect collision avoidance behavior for further analysis, we propose a domain transformation algorithm that detects it by transforming the trajectories in the spatial domain into a new domain where the behavior is much more apparent and is thus easier to detect. The new domain also provides a new perspective for understanding collision avoidance behavior. Second, since there are no mature metrics to evaluate the level of realism, we propose a new evaluation metric based on the least-effort theory, which evaluates the realism of collision avoidance behavior by its physical and mental consumption. This evaluation metric also provides the foundation of modeling. Third, for realistic crowd simulation, we design a reinforcement learning model. It trains agents with our proposed reward function that models pedestrians' intrinsic needs of "reducing effort consumption'' and thus can guide agents to behave realistically when avoiding collisions. Extensive experiments show our model is 55.9% and 52.5% more realistic in collision avoidance behavior than the best baselines on two real-world datasets. We release our codes at https://github.com/tsinghua-fib-lab/TECRL.

A Force-Based Model for Adaptively Controlling the Spatial Configuration of Pedestrian Subgroups at Non-Extreme Densities

Pedestrians’ behavior in emergency evacuation: Modeling and simulation*

Modifications to Social Force Model

Visually-Guided Pedestrian Crowd Simulation

A Local Behavior Model for Small Pedestrian Groups

Effects of Crowding Perception on Self-organized Pedestrian Flows Using Adaptive Agent-based Model

Modeling Dynamic Groups for Agent-Based Pedestrian Crowd Simulations

An Extended Social Force Model Via Pedestrian Heterogeneity Affecting the Self-Driven Force

Improvement Dynamic Continuum Model for Pedestrian Flow

Simulation Model of Self-Organizing Pedestrian Movement Considering Following Behavior.

A collective motion model based on two-layer relationship mechanism for bi-direction pedestrian flow simulation

A Modified Social Force Model Considering Relative Velocity of Pedestrians

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

Fusion-based Extended Social Force Model for Reciprocal Transformation Tasks in Bidirectional Pedestrian Movement

An all-densities pedestrian simulator based on a dynamic evaluation of the interpersonal distances

The Effect of Social Groups on the Dynamics of Bi-Directional Pedestrian Flow: A Numerical Study

Social Groups Barely Change the Speed-Density Relationship in Unidirectional Pedestrian Flow, but Affect Operational Behaviours

A Modified Heuristics-Based Model for Simulating Realistic Pedestrian Movement Behavior

Understanding and Modeling Collision Avoidance Behavior for Realistic Crowd Simulation

Modeling the Mutual Anticipation in Human Crowds with Attention Distractions

Sub-Goal Social Force Model for Collective Pedestrian Motion Under Vehicle Influence