Crowd behavior simulation based on shadow obstacle and ORCA models

Abstract

Crowd behavior simulation has become one of the key support technologies for evacuation drills and safety monitoring. However, complex psychology, personality traits and interactions among individuals increase the challenges of creating realistic crowd behavioral simulations. In this paper, we propose a crowd behavior simulation method to simulate crowd behaviors by combining the shadow obstacle model (SOM) and optimal reciprocal collision avoidance (ORCA) model. First, the SOM is converted to the expected speed, as well as the half plane of ORCA. Then, a manual generation method of SOM is proposed to simulate the crowd behaviors in corners or beside a non-closed wall. Finally, we combine behavior simulation and physical simulation techniques, and propose a velocity-based crowd simulation framework. To test the effectiveness of our proposed method, several simulation methods are implemented in a Unity3D engine-based crowd simulation in subway station environment, and the comparison results demonstrate that our proposed method outperforms force-based simulation methods.