Collision avoidance approach for example-based crowd simulation

Weiwei Xing,Lili Zhu,Xiang Wei,Peng Bao

DOI: https://doi.org/10.24507/ijicic.14.01.127

2017-01-01

Abstract:Collision avoidance plays an essential role in crowd simulation and determines the simulation quality to some extent. We propose a novel method to improve the collision avoidance performance based on the existing crowd simulation methods. The example-based method is chosen as the basic method by reasons of the realistic simulation result. Considering collision avoidance in the example-based method relies heavily on data captured from the real world, velocity obstacle (VO) is introduced to calculate an action of an agent when collision-free example cannot be found. We finally optimize the performance of collision avoidance by performing the novel collision fixing (CF) algorithm. The CF takes advantage of obtained actions to detect potential collisions and employs repulsive force to avoid the collisions. Experiments are conducted to evaluate the performance of the proposed method, and the results show that the method greatly reduces collision times, especially in complex simulation scenarios, while keeping the scenarios otherwise realistic.