一个可扩展的微观交通仿真平台的研究与实现

发布时间：2018-05-13 12:59

本文选题：微观交通仿真 + 路网模型　；参考：《浙江大学》2015年硕士论文

【摘要】：随着社会经济的发展,交通供需矛盾不断激化,已严重制约着城市的发展。由于道路交通系统是一个复杂的巨系统,单独的理论研究或实验方法都不能很好的胜任,而微观交通仿真方法以个体车辆为研究对象,模拟车辆与车辆、车辆与环境之间的相互作用,从而再现交通流的真实状态,展现交通控制策略的实施过程,成为解决这类现实问题的有效实验工具和探索方法。文章在研究学习国内外微观交通仿真现状的基础上,完整设计并实现了一个微观交通仿真平台,该平台面向高速公路和城市路网的交通仿真,具备高度的扩展性,可根据实际需求不同灵活增减功能。概括起来,平台的内容包含以下这些：首先,文章定义了路网的内存模型,通过Lane/Edge/Road三层结构来描述现实中复杂的路网结构和交通规则约束下的路网连通性状态。其次,文章实现了两个模型用来模拟驾驶员驾驶车辆的行为,其中跟车模型(CFM)控制车辆的纵向运动行为,换道模型(LCM)控制车辆的横向运动行为,两者分别解决车辆受其他车辆和环境影响下的车速变化和车道选择问题。然后,文章实现了一个微观交通模拟器,模拟器把某个时间点上整个交通系统的状态集合看作一帧,结合CFM和LCM返回的结果,模拟器并行、迭代计算每一帧,从而持续更新整个交通系统的状态。接着,文章实现了完整的车辆生命周期控制,共3个部分7种状态。除了车辆行驶模拟部分外,还包括：车辆初始化部分,完成车辆的创建和属性初始化,同时提出一种在交通规则约束下的最低代价路径计算算法；车辆插入与回收部分,实现高级的车辆进入、离开路网的控制逻辑。最后,作为一个平台,文章提供了多种接口和功能扩展机制,为了在提供扩展性的同时保证性能,文章还实现了一个专用的容器和一套高效的事件触发机制。
[Abstract]:With the development of social economy, the contradiction of traffic supply and demand is becoming more and more intense, which has restricted the development of the city. Because the road traffic system is a complicated giant system, the single theoretical research or experimental method can not do well. However, the microscopic traffic simulation method takes individual vehicles as the research object, and simulates vehicles and vehicles. The interaction between vehicle and environment can reproduce the real state of traffic flow and show the implementation process of traffic control strategy. It becomes an effective experimental tool and exploration method to solve these practical problems. On the basis of studying the present situation of micro traffic simulation at home and abroad, this paper designs and implements a micro traffic simulation platform, which is highly extensible for traffic simulation of freeway and urban road network. Can according to the actual demand different flexible increase or decrease function. To sum up, the contents of the platform include the following: firstly, this paper defines the memory model of the road network, and describes the complex road network structure and the road network connectivity behavior under the constraints of traffic rules through the Lane/Edge/Road three-layer structure. Secondly, two models are implemented to simulate the driver's driving behavior, in which the following model (CFM) controls the longitudinal motion of the vehicle, and the change of track model (LCM) controls the lateral movement of the vehicle. They solve the problem of vehicle speed change and lane selection under the influence of other vehicles and environment respectively. Then, a micro traffic simulator is implemented. The simulator regards the state set of the whole traffic system at a certain point as a frame. Combined with the results returned by CFM and LCM, the simulator computes each frame in parallel and iteratively. Thus continuously updating the state of the entire transportation system. Then, the paper realizes the complete vehicle life cycle control, which consists of 3 parts and 7 states. In addition to the vehicle driving simulation, it also includes the vehicle initialization part, which completes the vehicle creation and attribute initialization. At the same time, it proposes a minimum cost path calculation algorithm under the traffic rules constraint, and the vehicle insertion and recovery section. Achieve advanced vehicle entry and exit road network control logic. Finally, as a platform, the paper provides a variety of interfaces and function extension mechanisms. In order to provide scalability while ensuring performance, the paper also implements a special container and a set of efficient event trigger mechanism.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U491;TP391.9

【参考文献】