基于分层控制结构的迭代学习城市交通信号控制  

Hierarchical structure with iterative learning control for urban traffic networks

作　　者：杨曦[1] 黄青青 刘志[1] YANY Xi;HUANG Qingqing;LIU Zhi(College of Computer Science and Technology,Zhejiang University of Technology,Hangzhou 310023,China)

机构地区：[1]浙江工业大学计算机科学与技术学院,浙江杭州310023

出　　处：《浙江工业大学学报》2019年第3期305-311,共7页Journal of Zhejiang University of Technology

基　　金：国家自然科学基金资助项目(61603339);浙江省自然科学基金资助项目(LY16F020033)

摘　　要：分层控制是大规模城市路网实施高效信号控制的有效手段。根据城市路网规模庞大且结构复杂的特点,提出了一种基于分层控制结构、协调各子区交通状态的迭代学习城市交通信号控制策略。上层利用交通数据,刻画路网内各子区的宏观基本图(Macroscopic fundamental diagram, MFD),基于MFD分析得到子区车辆累积数与流量的关系,并以道路占有率均衡为目标,设计各子区理想的道路占有率;下层基于道路交通流模型,通过迭代学习获得各路口的信号配时方案,使子区内的道路占有率达到上层的要求。提出的分层控制策略使路网内交通流分布均衡,提高路网整体通行能力。Matlab和Vissim的仿真结果与Webster固定配时信号控制的对比显示了该控制策略的有效性和优越性。Hierarchical framework is essential for efficient control of large-scale urban traffic network.Under the complex traffic conditions,a two-level hierarchical signal control strategy is proposed to coordinate the traffic of subnetworks.At the upper level,the macroscopic fundamental diagrams (MFD) which reveal the relationship between the vehicle accumulation and outflow of the concerned subnetworks are derived,and the optimal balanced occupancy rate is provided for each subnetwork.At the lower level,based on the dynamic model of the traffic flow,the signal timing plan is designed to satisfy the optimal occupancy rate by using iterative learning technique.By using the aforementioned signal control strategy,the link densities are balanced within the whole network,and such feature of network homogeneity facilitates the improvement of traffic mobility and performances.The applicability of the proposed strategy is evaluated by Vissim and Matlab simulation,its efficiency is also provided by comparing with Webster fixed-time control.

关 键 词：分层控制 宏观基本图 迭代学习控制 信号控制 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]