行人交通微观仿真中碰撞检测模型研究

摘要

Collision detection is an inevitable importation issue in computer simulation application of pedestrian microscopic behavioral models. Based on the characteristics of pedestrian traffic combined with the practical need of computer simulation, this paper presented a new collision detection model suitable for microscopic simulation of pedestrian traffic. The model was based on the idea of "Speedbox" method, two pedestrian collision types "cross collision" and "fellow collision" with corresponding threshold algorithms. The model can solve the overlapping trouble in pedestrian microscopic simulation application, and help the results closer to the reality.%结合行人交通的特征和计算机仿真的实际需要,构建出一个新的适用于行人交通微观仿真的碰撞检测模型,该模型以经典的包围盒法为基本思路,分解出“交叉”和“同道”2类行人碰撞类型,并构建出了相应的阈值算法.该模型在实际应用中能较好地解决行人微观仿真中行人交叠的问题,使仿真效果更加贴近实际.