信号交叉口排队长度预测的神经网络方法

预测信号交叉口的排队长度可以为交通信号控制和管理提供非常重要的信息。基于神经网络，针对定时和感应信号交叉口两种不同情况，成功实现了单、双车道排队长度的预测。同时，感应器与停车线之间的距离对预测精度的影响也进行了初步分析。模拟结果同时表明，神经网络对左转车道排队长度的预测效果不佳，不能为信号控制和管理提供有效的信息。     

城市四交叉路口交通引导仿真系统

基于城市道路交通的拥挤和阻塞主要表现在若干个交叉路口的事实，提出了以四交叉路口为对象的交通引导仿真系统模型框架，并讨论了实际选取交叉路口应考虑的三个方面的因素。如果把每个交叉路口的车流看作排队事件，不同交叉路口的车流事件会发生冲突。轮询各个交叉路口的车流排队事件，用事件驱动状态转移与专家系统技术相结合的方法来疏解事件冲突，做出调整交通引导的决策。同时给出了仿真系统模型的目标、功能以及交叉口、车道、信号相位、信号周期等系统状态的描述。     

设置非机动车等候区的信号交叉口交通流模型

城市机非混行交叉口的管理与控制是交通管理的重要内容.针对信号交叉口设置非机动车等候区的交通组织方式，建立了交叉口元胞自动机模型.对机动车流采用基于经典 NaSch(NS)的改进多车道元胞自动机模型，建立了交叉口换道规则，增加主动减速规则；对非机动车流采用具有侧向运动的扩展多值 CA模型.研究了设置等候区的车流状态特性，以及非机动车密度和等候区纵向长度对信号交叉口的影响.研究结果表明：设置等候区能在一定程度上提高交叉口的通行能力，但长度并非越长越好，当长度过长时在一定条件下会增大对机动车流的阻滞，总体上等候区是一种值得借鉴的方式；非机动车密度对机动车流的基本图具有显著影响.     

基于Paramics的多相位感应信号控制仿真研究

介绍了微观交通仿真软件Paramics及其API函数,建立了基于微观交通仿真的感应信号控制策略技术路线,提出基于Paramics的感应信号控制仿真实现过程及相关技术,并结合仿真实例对仿真结果进行了分析.     

基于交叉口可靠性的公交优先信号配时优化模型

公交信号优先被广泛应用于提高道路通行能力和服务水平等方面,但是,公交信号优先经常会破坏路网,从而导致交叉口可靠性降低.因此,交叉口可靠性是交通信号设计中需考虑的一个重要问题.本研究针对单点交叉口提出了基于可靠性的公交信号优先配时优化模型,将人均延误最小化作为目标函数,将交叉口相位清空可靠度指标作为一个重要约束条件,该模型在优化信号配时的过程中不仅降低了人均延误,并且保证了交叉口的可靠性.本研究将该优化模型应用于单点交叉口,通过分析影响参数得到:仅提高可靠性要求,延误时间会非线性增加;若一味地降低总延误会导致低可靠性.最后,通过案例分析将传统模型与本研究模型进行对比分析,验证了本研究所提出模型的实用性.     

Traffic conflict models to evaluate the safety of signalized intersections at the cycle level

The safety of signalized intersections has often been evaluated at an aggregate level relating collisions to annual traffic volume and the geometric characteristics of the intersection. However, for many safety issues, it is essential to understand how changes in traffic parameters and signal control affect safety at the signal cycle level. This paper develops conflict-based safety performance functions (SPFs) for signalized intersections at the signal cycle level. Traffic video-data was recorded for six signalized intersections located in two cities in Canada. A video analysis procedure is proposed to collect rear-end conflicts and various traffic variables at each signal cycle from the recorded videos. The traffic variables include: traffic volume, maximum queue length, shock wave characteristics (e.g. shock wave speed and shock wave area), and the platoon ratio. The SPFs are developed using the generalized linear models (GLM) approach. The results show that all models have good fit and almost all the explanatory variables are statistically significant leading to better prediction of conflict occurrence beyond what can be expected from the traffic volume only. Furthermore, space-time conflict heat maps are developed to investigate the distribution of the traffic conflicts. The heat maps illustrate graphically the association between rear-end conflicts and various traffic parameters. The developed models can give insight about how changes in the signal cycle design affect the safety of signalized intersections. The overall goal is to use the developed models for the real-time optimization of signalized intersection safety by changing the signal design.     

Identifying the most critical transportation intersections using social network analysis

Traffic congestion negatively impacts our society. Most of the traditional transportation planning techniques – though effective – require rigorous amounts of data and analysis which consumes time and resources. This paper uses social network analysis (SNA) to analyze transportation networks, and consequently corroborate the effectiveness of SNA as a complementary tool for improved transportation planning. After creating the connection between the language and concepts of SNA and those of transportation systems – as well as developing a model that utilizes different SNA centrality measures within the transportation context – the authors utilize SNA to investigate traffic networks in three case studies in the state of Louisiana, analyze the results and draw conclusions. To this effect, with minimal cost and time, the model identifies the most critical intersections that should be further investigated using traditional techniques. These results are in agreement with the findings of Louisiana’s Department of Transportation and Development.     

非机动车影响下信号交叉口通行能力计算模型

为量化非机动车对信号交叉口通行能力的影响,分析了信号交叉口非机动车影响机动车运行的方式,估计了其持续时间,基于流量-速度关系,计算了不同情况下的饱和流率,建立了非机动车影响下典型信号交叉口通行能力模型。计算结果表明:利用本模型计算的左转、直行机动车通行能力总体上低于HCM(Highway Capacity Manual)计算值;而右转机动车通行能力计算值在非机动车流量较低时与HCM计算值接近,在非机动车流量较高时略高于HCM计算值。可见,此模型可有效应用于计算非机动车影响下的信号交叉口通行能力。     

四路信号控制交叉口交通事故预测模型研究

通过对北京郊区四路信号控制交叉口进行实地调查,在获取事故数据及道路几何要素、环境要素、交通流要素等大量数据的基础上,利用相关性分析寻找出与交通事故相关性显著的影响因素,并采用AIC原则分析法确定模型形式,分别建立了四路信号控制交叉口死亡人数、受伤人数、事故次数预测模型。模型的建立有助于对交叉口进行安全分析诊断及预测交叉口的交通事故发生规律。     

ӵ��״̬���źŽ����ͨ���������㷽��

车头时距分布是研究交通流特性的重要方法。本文基于北京市若干交叉口的实测样本数据,研究分析了拥堵状态下信号交叉口的排队车辆越过停车线的车头时距分布及其变化特点,结果表明在第6个车辆之后达到饱和车头时距。在此基础上,针对中心城区拥堵区域次干道以上等级的信号交叉口,提出拥堵状态下的通行能力模型。进一步地,利用实测数据作了案例计算与对比分析,直行与左转车道通行能力的平均误差率分别为2.27%与4.28%,验证了模型的可行性与有效性。此外,该结果优于其它计算方法所得结果,表明在拥堵状态下,该方法的适用性更高。