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.

Visually-Guided Pedestrian Crowd Simulation

Modeling and Simulating the Walking Behavior of Small Pedestrian Groups.

Within-crowd immersive evaluation of collision avoidance behaviors.

Optimal Acceleration-Velocity-Bounded Trajectory Planning in Dynamic Crowd Simulation.

Improvement Dynamic Continuum Model for Pedestrian Flow

A Simple and Realistic Pedestrian Model for Crowd Simulation and Application

Group Modeling: A Unified Velocity-Based Approach.

Understanding and Modeling Collision Avoidance Behavior for Realistic Crowd Simulation

A Generic Bi-Layer Data-Driven Crowd Behaviors Modeling Approach

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

Inserting Virtual Pedestrians into Pedestrian Groups Video with Behavior Consistency

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

A Hybrid Model for Simulation of Crowd Evacuation

Simulation of Virtual Crowd Behaviors in Emergency

Collision Avoidance in an Agent System for Crowd Simulation

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

Simple Cognitive Heuristics Applied to Modeling Pedestrian Behavior Dynamics

A Local Behavior Model for Small Pedestrian Groups

A Hybrid Pedestrian Motion Modeling Approach for Crowd Simulation

Data-Driven and Collision-Free Hybrid Crowd Simulation Model for Real Scenario

Field-based crowd simulation