多障碍物场景中的行人运动模型

Pedestrian model combined with detour behavior in multi-obstacles environment

社会力模型是一类重要的行人运动模型,在行人建模仿真领域得到广泛应用.然而该模型仅能应用于简单空间场景,当空间中有障碍物,行人与目标点不满足通视条件时,行人不能绕过障碍,会出现穿透障碍、轨迹振荡等问题.本文引入规避障碍行为,与社会力模型融合在一起,实现行人在复杂二维空间中的运动.规避障碍行为的核心是确定绕行点,以期望行走路线最短为准则构造了求解绕行点的计算几何方法,并解决复杂空间场景中的多重遮挡问题.建立行人运动逻辑,设计渐进式路径规划算法,实现行人运动.开发了行人仿真系统,对所提出的模型进行验证.仿真结果表明：在多障碍物空间场景中,行人能自然的规避障碍,形成合理的运动轨迹.

The social force model is widely used in the field of pedestrian modeling and simulation. However, it can only be applied to simple scenarios. When there are many obstacles in the environment, and the pedestrian cannot see the goal, the model will lead to problems such as obstacle penetrating and trajectory oscillation. We proposed a hybrid model composed of tactical layer and operation layer. The core concept of the tactical layer is detour point. We constructed a computational geometry method to find the detour point based on the criterion of shortest path, and solved the multiple occlusion problem. Then we designed a progressive route planning algorithm to guild pedestrian＇s movement. Finally, a simulation system is developed to verify the proposed model. The simulation results show that, both in single obstacle and multiple obstacle environments, pedestrian can avoid the obstacles naturally and the trajectory is realistic and reasonable.