考虑同伴群的行人感知避让社会力改进模型
发布时间：2019-05-22 05:32
【摘要】：为了提高人群行为模拟的真实性,针对现有社会力模型效果展现存在的行人间避让效果不佳以及个体、同伴群成员差异性避让行为刻画不足的问题,引入融入Agent的实时感知与动态避让算法对模型进行改进研究。基于面向对象的编程语言VC++,结合应用程序框架QT进行面向对象软件的设计与开发,最终搭建了行人仿真平台。通过行人运动状态参数的标定,实现了模型对行人步行行为更加真实的模拟,并进行了不同控制条件下单、双向行人流仿真试验。结果表明,改进模型能够更加真实地模拟冲突行人之间的避让,同伴群成员别于个体进行避让的行为特征得以体现;单向行人流的通行效率远高于双向行人流;避让行为的引入减缓了行人碰撞前后的速度变化,但降低了行人流的整体运动速度。
[Abstract]:In order to improve the authenticity of crowd behavior simulation, aiming at the problem that the effect of existing social force model shows that the pedestrian avoidance effect is not good and the difference avoidance behavior of individuals and peer group members is insufficient, The real-time perception and dynamic avoidance algorithm integrated with Agent is introduced to improve the model. Based on the object-oriented programming language VC and the application framework QT, the object-oriented software is designed and developed, and finally the pedestrian simulation platform is built. Through the calibration of pedestrian motion state parameters, the model is used to simulate the pedestrian behavior more realistically, and the orders and bidirectional pedestrian flow simulation experiments are carried out under different control conditions. The results show that the improved model can simulate the avoidance between collision pedestrians more realistically, and the behavior characteristics of peer group members are different from those of individuals, and the traffic efficiency of one-way pedestrian flow is much higher than that of two-way pedestrian flow. The introduction of avoidance behavior slows down the speed change before and after pedestrian collision, but reduces the overall movement speed of pedestrian flow.
【作者单位】： 昆明理工大学交通工程学院;昆明市规划设计研究院;昆明轨道交通集团有限公司;洛阳市规划建筑设计研究院有限公司;
【基金】：国家自然科学基金项目(51168023)
【分类号】：U491
，
本文编号：2482719
[Abstract]:In order to improve the authenticity of crowd behavior simulation, aiming at the problem that the effect of existing social force model shows that the pedestrian avoidance effect is not good and the difference avoidance behavior of individuals and peer group members is insufficient, The real-time perception and dynamic avoidance algorithm integrated with Agent is introduced to improve the model. Based on the object-oriented programming language VC and the application framework QT, the object-oriented software is designed and developed, and finally the pedestrian simulation platform is built. Through the calibration of pedestrian motion state parameters, the model is used to simulate the pedestrian behavior more realistically, and the orders and bidirectional pedestrian flow simulation experiments are carried out under different control conditions. The results show that the improved model can simulate the avoidance between collision pedestrians more realistically, and the behavior characteristics of peer group members are different from those of individuals, and the traffic efficiency of one-way pedestrian flow is much higher than that of two-way pedestrian flow. The introduction of avoidance behavior slows down the speed change before and after pedestrian collision, but reduces the overall movement speed of pedestrian flow.
【作者单位】： 昆明理工大学交通工程学院;昆明市规划设计研究院;昆明轨道交通集团有限公司;洛阳市规划建筑设计研究院有限公司;
【基金】：国家自然科学基金项目(51168023)
【分类号】：U491
，
本文编号：2482719
本文链接：https://www.wllwen.com/kejilunwen/daoluqiaoliang/2482719.html
