基于视频数据的交叉口状态判别及排队长度估计  

State Discriminant and Queue Length Estimation of an Intersection Based on Video Data

作　　者：唐少虎[1] 刘小明[1] 陈兆盟[1] 

机构地区：[1]北方工业大学城市道路交通智能控制技术北京市重点实验室,北京100144

出　　处：《道路交通与安全》2015年第1期58-64,共7页Road Traffic & Safety

基　　金：国家自然科学基金(61374191);国家高技术研究发展计划(863计划)(2012AA112401)

摘　　要：为了评价信号控制交叉口的交通状态及估计周期排队长度,本文以集散波为理论基础,分析视频检测的机动车通过上下游相邻交叉口的数据,提出了基于车辆延误时间的交叉口交通状态判别划分,进一步设计了欠饱和及饱和交通状态下的交叉口最大排队长度估计方法.实例分析表明,本文方法所估计的排队长度比较接近实际调查值,排队相对误差在可接受的范围内,结果验证了所设计的排队长度估计方法是有效可行的.In order to estimate traffic states and periodic queue length evaluation of signal control intersection, this paper analyzed video detection data of vehicles by passing through adjacent intersections,proposed traffic state discriminant of an intersection based on the vehicles delay time,and developed the maximum queue length estimation method of an intersection in under-saturation or saturation traffic condition. The empirical analysis showed that the estimated queue length by the proposed method was closer to the actual investigation,and queuing relative error was in an acceptable range. The results validated that the designed method of queue length estimation of an intersection is feasible and effective.

关 键 词：视频数据 集散波 交通状态判别 最大排队长度估计 

分 类 号：U491[交通运输工程—交通运输规划与管理]