基于视频信息的单点瓶颈控制策略与实现方法

项目名称： 基于视频信息的单点瓶颈控制策略与实现方法

项目编号： No.61304191

项目类型： 青年科学基金项目

立项/批准年度： 2014

项目学科： 自动化技术、计算机技术

项目作者： 马东方

作者单位： 浙江大学

项目金额： 23万元

中文摘要： 基于断面交通数据的传统信号控制技术及控制系统在信息采集与算法设计方面主要面向于未饱和交通状态，而过饱和控制方法的实践效果不佳。随着闯红灯自动记录系统等视频设备的广泛应用，使得通过深入挖掘视频信息进而提取行程时间、节点OD等区间交通参数已成为可能，可为过饱和状态下的瓶颈控制提供数据基础。本项目基于日益普及的视频检测设备与日益成熟的视频处理技术，从信息提取、信息利用两个层面入手，研究基于视频信息的断面及区间交通流数据获取方法，并依据视频信息的时空关联特性确定数据处理间隔，建立交通流参数的预测与校正模型；构建交通状态指标集，通过影响因素灵敏性分析优选视频信息下的状态表达指标，进而建立路段交通状态动态演化模型并界定瓶颈触发阈值；利用区间交通信息，从相位相序与配时参数综合优化的角度寻求快速消散单点"瓶颈"的信号控制方法。项目成果可为预防短连线排队溢流、减少交通拥堵时空扩散提供理论支撑与实现方法。

中文关键词： 视频检测；行程时间；瓶颈状态；瓶颈识别；瓶颈控制

英文摘要： The implementation effect of traditional signal control technologies and mainstream systems is not good under oversaturation conditions, as the information collection method and the core algorithms in these technologies and systems are mainly applicable to under-saturation state. Meanwhile, RLRARS (Red-light-running-automated-recording-system) can accurately record the information of vehicles that passing through the virtual testing interval, and we could obtain the traffic parameters, such as traffic flow, link travel time and node OD under high saturation condition by analyzing the detection data. This project will carry out on the basis of that the video detection technology is increasing maturity and the devices are growing popularity, and the task can be divided into two parts: information extraction and utilization. The first part will focus on the method to extract the traffic parameters, especially the travel time and node OD from RLRARS, and then develop the prediction and correction models for these parameters after analyzing the spatial and temporal correlation characteristics of the video information and determining the data processing interval. Secondly, we will optimize the index for road state by significance analysis of influence factors, and advance a description method for road state evolution

英文关键词： Video-imaging detecting；Travel time；Bottleneck；Bottleneck identification；Bottleneck control