面向交通控制的时段划分与子区划分
发布时间：2018-10-24 13:54
【摘要】：交通流的波动性和不确定性是交通控制复杂性的主要原因之一。这种波动性不仅表现在不同时间上的交通流变化,也表现在不同空间上的交通流差异,本文主要研究面向交通控制的时段划分与子区划分,其中子区划分部分又可根据控制区域的大小,分为用于干线协调控制的协调路口划分和用于区域协调控制的子区划分。 在时段划分方面,分析了聚类数据的组成和间隔对时段划分的影响,在对经典的NJw谱聚类算法进行改进后,应用在一个路口和多个路口的时段划分中,并与K-means算法得到的结果进行仿真对比.然后考虑使用多天的交通流数据后,对时段划分流程进行改进,使用主成分分析法对数据进行降维处理后,使用Self-Tuning谱聚类算法进行时段划分,从聚类效果和仿真试验两方面证明了该方法的有效性。接着考虑到交通流的时序特征,使用有序聚类对最佳周期数据进行分析,效果优于K-means算法。 在协调路口划分方面,首先分析了Synchro软件中用于描述路口关联性的协调因子计算模型,然后用实例分析了Synchro软件划分协调路口的步骤和方法,仿真结果表明Synchro软件无法直接得到最优的划分,而且协调路口的划分对协调控制的效果具有显著的影响。然后根据通过路口数最大的模型,建立了协调路口的划分标准,方法简单有效。 在子区划分方面,首先微观仿真软件VISSIM建立仿真环境,分别分析一个路口周期、绿信比、饱和度的变化对邻近路口控制方案和控制效果的影响,以深入理解交叉口的关联性。然后利用复杂网络中的社团划分算法,进行改进后应用在加权网络中,得到大范围交通网络的子区划分方法。
[Abstract]:The fluctuation and uncertainty of traffic flow is one of the main reasons for the complexity of traffic control. This fluctuation is not only reflected in the traffic flow variation in different time, but also in the traffic flow difference in different space. In this paper, the time division and sub-area division for traffic control are mainly studied. According to the size of the control area, the sub-area partition can be divided into coordinated intersection division for trunk line coordination control and sub-area division for area coordination control. In the aspect of time division, the influence of the composition and interval of clustering data on time division is analyzed. After improving the classical NJw spectral clustering algorithm, it is applied to the time division of one or more intersections. The simulation results are compared with the results obtained by K-means algorithm. Then, after considering the use of multi-day traffic flow data, the process of time division is improved. After dimensionality reduction of data is processed by principal component analysis, Self-Tuning spectral clustering algorithm is used for time division. The effectiveness of the method is proved by clustering effect and simulation experiment. Then considering the time series characteristics of traffic flow, the best periodic data is analyzed by order clustering, and the result is better than that of K-means algorithm. In the aspect of coordinating intersection division, this paper first analyzes the calculation model of coordination factor used in Synchro software to describe the correlation of intersection, and then analyzes the steps and methods of dividing coordinated intersection with Synchro software with an example. The simulation results show that the optimal partition can not be obtained directly by Synchro software, and the partition of coordinated intersection has a significant effect on the effect of coordinated control. Then, according to the model of maximum number of intersections, the dividing standard of coordinated intersection is established. The method is simple and effective. In sub-area division, the simulation environment is established by micro-simulation software VISSIM, and the influence of the change of intersection cycle, green signal ratio and saturation on the control scheme and control effect of adjacent intersection is analyzed respectively, in order to understand the correlation of intersection deeply. Then using the community partition algorithm in the complex network, the improved algorithm is applied to the weighted network, and the sub-area partition method of the large-scale traffic network is obtained.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U491.54
本文编号：2291597
[Abstract]:The fluctuation and uncertainty of traffic flow is one of the main reasons for the complexity of traffic control. This fluctuation is not only reflected in the traffic flow variation in different time, but also in the traffic flow difference in different space. In this paper, the time division and sub-area division for traffic control are mainly studied. According to the size of the control area, the sub-area partition can be divided into coordinated intersection division for trunk line coordination control and sub-area division for area coordination control. In the aspect of time division, the influence of the composition and interval of clustering data on time division is analyzed. After improving the classical NJw spectral clustering algorithm, it is applied to the time division of one or more intersections. The simulation results are compared with the results obtained by K-means algorithm. Then, after considering the use of multi-day traffic flow data, the process of time division is improved. After dimensionality reduction of data is processed by principal component analysis, Self-Tuning spectral clustering algorithm is used for time division. The effectiveness of the method is proved by clustering effect and simulation experiment. Then considering the time series characteristics of traffic flow, the best periodic data is analyzed by order clustering, and the result is better than that of K-means algorithm. In the aspect of coordinating intersection division, this paper first analyzes the calculation model of coordination factor used in Synchro software to describe the correlation of intersection, and then analyzes the steps and methods of dividing coordinated intersection with Synchro software with an example. The simulation results show that the optimal partition can not be obtained directly by Synchro software, and the partition of coordinated intersection has a significant effect on the effect of coordinated control. Then, according to the model of maximum number of intersections, the dividing standard of coordinated intersection is established. The method is simple and effective. In sub-area division, the simulation environment is established by micro-simulation software VISSIM, and the influence of the change of intersection cycle, green signal ratio and saturation on the control scheme and control effect of adjacent intersection is analyzed respectively, in order to understand the correlation of intersection deeply. Then using the community partition algorithm in the complex network, the improved algorithm is applied to the weighted network, and the sub-area partition method of the large-scale traffic network is obtained.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U491.54
【参考文献】
相关期刊论文 前10条
1 邹志云,陈绍宽,郭谨一,白立琼,常超凡;基于Synchro系统的典型信号交叉口配时优化研究[J];北京交通大学学报;2004年06期
2 胡华;高云峰;邱薇华;;基于路径关联性的交叉口群动态划分方法[J];重庆交通大学学报(自然科学版);2010年05期
3 宫晓燕,陈德望,刘小明;基于有序样本聚类的交叉口交通信号控制算法研究[J];公路交通科技;2003年01期
4 孙超;徐建闽;;基于Synchro的单点交叉口信号配时优化研究[J];公路交通科技;2009年11期
5 谢颖;李春贵;闫向磊;张宏磊;;基于模糊神经网络交通控制子区动态划分[J];广西工学院学报;2011年01期
6 卢凯;徐建闽;李轶舜;;基于关联度分析的协调控制子区划分方法[J];华南理工大学学报(自然科学版);2009年07期
7 王殿海;徐程;祁宏生;金盛;;基于路段流量相关性的检测器优化布设[J];华南理工大学学报(自然科学版);2011年03期
8 杨庆芳;陈林;;交通控制子区动态划分方法[J];吉林大学学报(工学版);2006年S2期
9 段后利;李志恒;张毅;胡坚明;;交通控制子区动态划分模型[J];吉林大学学报(工学版);2009年S2期
10 董春娇;邵春福;陈晓明;李娟;;基于交通流时空特性的道路网动态划分方法[J];吉林大学学报(工学版);2010年06期
，本文编号：2291597
本文链接：https://www.wllwen.com/kejilunwen/jiaotonggongchenglunwen/2291597.html
