考虑车流异质性的信号交叉口疏解车辆跟驰模型

为客观地描述绿灯期间交叉口进口道异质疏解车流的跟驰行为,基于实测数据验证全速差模型发现,其加速度、速度、车间距的仿真结果存在较大误差. 考虑不同车型车辆性能和驾驶员驾驶行为差异,基于4 种跟驰情景,即小客车跟驰小客车(car-car),小客车跟驰公交车(car-bus),公交车跟驰小客车(bus-car),公交车跟驰公交车(bus-bus),建立考虑车流异质性的车辆跟驰模型. 结果表明,改进模型的性能提升明显,较全速差模型,速度和跟驰间距的均方根百分比误差(RMSPE)分别下降了15.29%,22.32%,更符合交叉口进口道异质疏解车流的跟驰行为.     

无人驾驶车辆城市交叉口周边车辆轨迹预测

针对城市交叉口周边车辆长时轨迹预测问题,搭建路基和实车采集平台采集大量轨迹数据,采用高斯混合模型(Gaussian Mixture Model,GMM)识别目标运动模式,采用高斯过程回归(Gaussian Processes Regression,GPR)模型进行城市交叉口周边车辆轨迹长时预测,采用路基数据集对预测模型...     

信号控制平面交叉口通行能力与信号周期之间的关系

探讨了信号控制交叉口的通行能力与信号周期之间的关系,研究发现信号交叉口进口道的通行能力并不一定随信号周期之增加而增加;给出进口道通行能力与信号周期之间关系的判断准则,并运用该准则进一步分析整个交叉口的通行能力与信号周期的关系。     

信号交叉口计算机仿真系统评价指标研究

对交叉口车辆运行评价指标建立原则的分析表明,建立相对独立有效的交叉口仿真系统运行评价指标,同时在对各评价指标的仿真属性分析的基础上,提出系统中评价指标的计算方法与具体实现原理,具有重要的现实意义.     

基于模糊时序网络的城市路网关键交叉口识别方法

在城市路网拓扑结构和动力学过程的基础上，增加对其时序特性的考虑，提出适用于城市路网关键交叉口识别的模糊时序网络模型。首先，阐述一般时序网络的描述方法和超邻接矩阵时序网络模型的原理，分析其优势以及将其用于城市路网分析的局限性；然后，提出优化措施，一 方面结合交通网络的功能特性，以动态交通参数构造单个时间步网络的层内交叉口交互强度模糊指标，另一方面借鉴并改进邻居拓扑重叠系数，对其进行模糊化处理，实现两相邻时间步网络层间交叉口关联强度的差异化表达；之后，在改进时间步层内、层间关联描述矩阵基础上，搭建模糊超邻接矩阵(Fuzzy Supra-adjacency Matrix, FSAM)时序网络模型(FSAM模型)；最后，以某城市核心区域147个交叉口构成的路网数据验证模型有效性。结果表明：以时序网络模型分析交叉口重要性非常必要，以中位数表达交叉口在时段内的重要性排序更为可靠；FSAM模型对交叉口重要性的排名时间序列有阶段持续性特征，且相比于特定时间步下基于单一指标的关键交叉口识 别结果具有更丰富的内涵；不同时间颗粒度下，FSAM模型对交叉口重要性排序的一致性较好，结果较为稳定。综上，该模型可供城市路网关键交叉口识别之用。     

Connected vehicle penetration rate for estimation of arterial measures of effectiveness

The Connected Vehicle (CV) technology is a mobile platform that enables a new dimension of data exchange among vehicles and between vehicles and infrastructure. This data source could improve the estimation of Measures of Effectiveness (MOEs) for traffic operations in real-time, allowing to perfectly monitor traffic states after being fully adopted. However, as with any novel technology, the CV adoption will be a gradual process. This research focuses on determining minimum CV technology penetration rates that would guarantee accurate MOE estimates on signalized arterials. First, we present estimation methods for various MOEs such as average speed, number of stops, acceleration noise, and delay, followed by an initial assessment of the penetration rates required to accurately estimate them in undersaturated and oversaturated conditions. Next, we propose a methodology to determine the minimum CV market penetration rates to guarantee accurate MOE estimates as a function of traffic conditions, signal settings, sampling duration, and the MOE variability. A correction factor is also provided to account for small vehicle populations where sampling is done without replacement. The methodology is tested in a simulated segment of the San Pablo Avenue arterial in Berkeley, CA. The outcomes show that the minimum penetration rate required can be estimated within 1% for most MOEs under a wide range of traffic conditions. The proposed methodology can be used to determine if MOE estimates obtained with a portion of CV equipped vehicles can yield accurate enough results. The methodology could also be used to develop and assess control strategies towards improved arterial traffic operations.     

Optimal operation of displaced left-turn intersections: A lane-based approach

Displaced left-turn (DLT) intersections that resolve the conflict between left-turn and opposing-through movements at the pre-signal are probably the most extensively used innovative intersection designs. The DLT intersection concept can be extended to ten different types according to the location of the left-turn transition area, the number of DLT approaches, and the possible setting of the bypass right-turn lane. This paper presents a generalized lane-based optimization model for the integrated design of DLT intersection types, lane markings, the length of the displaced left-turn lane, and the signal timings. The optimization is formulated as a mixed-integer non-linear program. This program is further transformed to a series of mixed-integer linear programming problems that can be solved by the standard branch-and-bound technique. Results from extensive numerical analyses reveal the effectiveness of the proposed method, as well as the promising property of assisting transportation professionals in the proper selection of DLT intersection types, and the design of geometric layout and signal timings.     

基于混行跟驰仿真的CAV专用进口道动态设置方法

信号交叉口是影响交通系统运行安全和效率的关键。在国家新基建战略的提出以及车路协同技术不断发展的环境下,合理设置网联自动驾驶车辆(Connected and Autonomous Vehicle,CAV)专用进口道,对信号交叉口进口道处不同网联类型的车辆进行科学的交通组织,能够提高交叉口的通行能力,降低行车延误,促进城市交通系统效率与安全的双提升。建立协同自适应巡航控制(Cooperative Adaptive Cruise Control,CACC)车辆跟驰模型和GM (General Motor)模型分别描述混行环境下网联车辆与非网联车的跟驰行为,以提高进口道通行能力、降低延误和油耗为优化目标,采取敏感度分析方法,提出不同CAV比例、进口道车道数、交通量和信号配时方案组合情况下CAV专用进口道的动态设置条件,适用于不同交通状况的信号交叉口,具有较强的普适性。数值仿真结果表明：采用该方法设置CAV专用进口道能够提高混行信号交叉口的通行能力、降低延误和车均油耗;在实际应用时,可视交叉口类型和交通智能化程度灵活选取CAV专用进口道设置方式,为混行交通流环境下交叉口进口道的交通组织优化提供理论依据和模型支持,对车路协同系统的相关研究具有参考意义。     

短连线交叉口群通行能力计算方法研究

针对短连线交叉口群的通行能力计算提出了相应的理论和方法。文章首先从不同角度对短连线交叉口群进行了界定，然后利用波动理论和车流运行的时空图，建立了短连线通行能力和损失的通行能力计算模型，并采用数学方法建立了通行能力及其损失量随着各个影响因素变化的图表。从中得出了一些重要的结论，且在信号协调控制方面提出了一些建议。