城市道路区域交通信号控制的动态子区划分
发布时间：2018-12-29 09:00
【摘要】：区域交通信号控制的子区划分是对大规模城市路网执行区域信号控制的基础。由于路网内交通流的复杂性以及时变特性,固定不变的子区方案大大降低了信号控制系统的实时应变能力。因此,对控制区域实行动态的子区划分逐渐成为研究热点。针对目前子区划分指标选取和划分方法存在的不足,本文提出了一种基于自组织映射神经网络算法的子区划分方法,结合划分过程中与关联性相关的动态指标的选取以及关键交叉口的实时判别,实现了控制区域的动态划分。在城市路网交通流运行特性分析方面,论文基于城市道路交叉口与路段的实测交通流数据,对路网内交通流的运行特性进行了分析。路网交通需求的传递是以交通流在时间和空间上的传播作为载体,本文采用交通需求在空间上的传递特性表征区域之间存在的关联特性。此外,论文介绍了宏观基本图的原理,并以此来研究路网的宏观交通特性。城市路网交通流运行特性研究为动态子区的划分奠定了理论基础。在区域信号控制的动态子区划分方面,论文主要针对动态子区划分所涉及的交叉口关联性影响因素和划分计算方法进行了深入的研究。通过对控制区域内关键交叉口的实时判别、影响交叉口之间关联特性因素的仿真分析以及自组织映射神经网络的计算原理等动态子区划分问题的研究,本文提出了一种基于自组织映射神经网络算法的动态子区划分方法。本文选取昆山市中心城区为研究对象,对基于自组织映射神经网络算法的动态子区划分方法效果进行了评估。论文采用基于自组织映射神经网络算法的动态子区划分方法和基于“合并指数”的动态子区划分方法对昆山市的路网划分进行演算。并采用基于路网空间交通分布和基于区域信号控制效益两个方面分别对两种划分方法进行评价。评估结果显示,本文所提出的动态子区划分方法能够很好将交通相似性高的交叉口和路段划分到一个控制子区,且对划分的控制子区执行协调控制时能够取得良好的控制效益。
[Abstract]:The subdivision of regional traffic signal control is the basis of regional signal control for large-scale urban road network. Because of the complexity and time-varying characteristics of traffic flow in the road network, the fixed sub-area scheme greatly reduces the real-time strain capacity of the signal control system. Therefore, the dynamic subdivision of the control area has gradually become a research hotspot. In view of the shortcomings of the current subdivision index selection and partition method, this paper proposes a subzone partition method based on self-organizing mapping neural network algorithm. The dynamic division of the control area is realized by combining the selection of correlation related dynamic indexes and the real-time identification of key intersections in the process of partition. In the aspect of analysis of traffic flow characteristics of urban road network, based on the measured traffic flow data of intersection and section of urban road, the operation characteristics of traffic flow in urban road network are analyzed in this paper. The transport of road network traffic demand is based on the transport flow in time and space as the carrier. This paper uses the transport demand in the space to express the correlation between the regions. In addition, the principle of macroscopic basic map is introduced, and the macroscopic traffic characteristic of road network is studied. The study of traffic flow characteristics of urban road network lays a theoretical foundation for the division of dynamic subareas. In the area of dynamic sub-area division of regional signal control, this paper mainly focuses on the influence factors of intersection correlation and the calculation method of dynamic sub-area division. Through the real-time identification of key intersections in the control area, the simulation analysis of the factors affecting the intersections' correlation characteristics, and the study of the dynamic sub-area partition problem, such as the calculation principle of self-organizing mapping neural networks, etc. In this paper, a dynamic subregion partition method based on self-organizing mapping neural network algorithm is proposed. In this paper, the effect of dynamic subarea partition method based on self-organizing mapping neural network algorithm is evaluated. In this paper, the dynamic subarea partition method based on self-organizing mapping neural network algorithm and the dynamic sub-area partition method based on "merge index" are used to calculate the road network division in Kunshan City. The two partition methods are evaluated based on the spatial traffic distribution of road network and the benefit of regional signal control respectively. The evaluation results show that the proposed method can well divide the intersection and section with high traffic similarity into a control sub-area, and can achieve good control efficiency when performing coordinated control for the divided control sub-area.
【学位授予单位】：东南大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：U491.54
本文编号：2394625
[Abstract]:The subdivision of regional traffic signal control is the basis of regional signal control for large-scale urban road network. Because of the complexity and time-varying characteristics of traffic flow in the road network, the fixed sub-area scheme greatly reduces the real-time strain capacity of the signal control system. Therefore, the dynamic subdivision of the control area has gradually become a research hotspot. In view of the shortcomings of the current subdivision index selection and partition method, this paper proposes a subzone partition method based on self-organizing mapping neural network algorithm. The dynamic division of the control area is realized by combining the selection of correlation related dynamic indexes and the real-time identification of key intersections in the process of partition. In the aspect of analysis of traffic flow characteristics of urban road network, based on the measured traffic flow data of intersection and section of urban road, the operation characteristics of traffic flow in urban road network are analyzed in this paper. The transport of road network traffic demand is based on the transport flow in time and space as the carrier. This paper uses the transport demand in the space to express the correlation between the regions. In addition, the principle of macroscopic basic map is introduced, and the macroscopic traffic characteristic of road network is studied. The study of traffic flow characteristics of urban road network lays a theoretical foundation for the division of dynamic subareas. In the area of dynamic sub-area division of regional signal control, this paper mainly focuses on the influence factors of intersection correlation and the calculation method of dynamic sub-area division. Through the real-time identification of key intersections in the control area, the simulation analysis of the factors affecting the intersections' correlation characteristics, and the study of the dynamic sub-area partition problem, such as the calculation principle of self-organizing mapping neural networks, etc. In this paper, a dynamic subregion partition method based on self-organizing mapping neural network algorithm is proposed. In this paper, the effect of dynamic subarea partition method based on self-organizing mapping neural network algorithm is evaluated. In this paper, the dynamic subarea partition method based on self-organizing mapping neural network algorithm and the dynamic sub-area partition method based on "merge index" are used to calculate the road network division in Kunshan City. The two partition methods are evaluated based on the spatial traffic distribution of road network and the benefit of regional signal control respectively. The evaluation results show that the proposed method can well divide the intersection and section with high traffic similarity into a control sub-area, and can achieve good control efficiency when performing coordinated control for the divided control sub-area.
【学位授予单位】：东南大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：U491.54
【参考文献】
相关期刊论文 前10条
1 冯远静;单敏;乐浩成;张贵军;俞立;;绿波协调控制的子区动态划分算法[J];控制理论与应用;2014年08期
2 别一鸣;卢凯;王琳虹;;相邻交叉口周期时长差异对关联度指标的影响[J];北京工业大学学报;2013年11期
3 Zhao ZHOU;Shu LIN;Yugeng XI;;A fast network partition method for large-scale urban traffic networks[J];Journal of Control Theory and Applications;2013年03期
4 首艳芳;徐建闽;;基于群体动力学的协调控制子区划分[J];华南理工大学学报(自然科学版);2013年04期
5 赵文涛;钱晓杰;朱芸;宋驰;沈国江;;基于关键路口的动态交通子区划分[J];计算机与应用化学;2012年10期
6 卢凯;徐建闽;郑淑鉴;王世明;;协调控制子区快速动态划分方法研究[J];自动化学报;2012年02期
7 马莹莹;杨晓光;曾滢;;基于谱方法的城市交通信号控制网络小区划分方法[J];系统工程理论与实践;2010年12期
8 杨晓光;黄玮;马万经;;过饱和状态下交通控制小区动态划分方法[J];同济大学学报(自然科学版);2010年10期
9 马万经;李晓丹;杨晓光;;基于路径的信号控制交叉口关联度计算模型[J];同济大学学报(自然科学版);2009年11期
10 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
相关硕士学位论文 前2条
1 李敏;区域边界交通控制及诱导协同技术研究[D];北方工业大学;2014年
2 魏勇;城市区域交通信号控制及交通状态分析研究[D];浙江大学;2013年
，本文编号：2394625
本文链接：https://www.wllwen.com/kejilunwen/jiaotonggongchenglunwen/2394625.html
