由于城市快速路出口匝道处驶出车流量过大、换道行为频繁，出口匝道处的排队往往蔓延至主线，致使主线通行能力下降、交通拥挤的现象日益突出。速度协调控制技术作为常用的主动式控制策略，能否缓解交通拥堵、提高交通运行效率成为近年来的研究热点。本文在现有速度协调控制模型的基础上，结合出口匝道瓶颈附近主线的交通流运行特征，提出基于车道的速度协调控制策略，从主线上寻求缓解出口匝道瓶颈附近交通拥堵的方法。

       首先，本文以Gungong Road立交出口匝道常发性瓶颈为例，采用基于车道的速度变形累计曲线法对瓶颈附近主线各个车道的交通流失效进行识别，从而可以有效判别半拥挤交通状态、详细分析交通拥堵的横向纵向传播规律，之后进一步比较瓶颈区域各车道交通拥挤前后的交通流率变化特征和内外车道的流量-速度关系图。研究表明，Gungong Road立交出口匝道瓶颈附近主线内外车道的交通流特征存在较大差异，因此以车道为对象进行精细化管理具有重要意义。

       之后，在现有基于断面的速度协调控制模型的基础上，本文通过对入口匝道、出口匝道等处的修正，采用METANET交通流模型描述和预测每个车道的交通流状态，以路网行程时间最短、总通过量最大为优化目标，构建基于车道的速度协调控制模型，对每个车道分别进行最佳速度引导。

       最后，本文利用VISSIM仿真软件、MATLAB软件、Visual Basic软件搭建Gungong Road立交出口匝道瓶颈的仿真系统，分析不同交通需求情形下，基于车道和基于断面的速度协调控制策略的实施效果。仿真结果表明：速度协调控制策略更加适用于中高密度，当交通需求量较高时，相比于无控制策略，基于车道和基于断面的速度协调控制模型均能降低车辆平均延误和停车次数，对缓解交通拥堵、提高交通安全具有一定助益，且在大多数交通需求情形下，基于车道的速度协调控制模型在车均延误、拥堵时空范围、停车次数等方面均优于基于断面的控制模型。为此进一步对基于车道的速度协调控制策略的实施形式进行初步探讨。

   Because of massive departing vehicles and frequent lane-changing behaviors near off-ramp on urban expressway, the queue at off-ramp often spread to the main line which would cause the capacity decline and traffic congestion on main line. And this phenomenon is becoming increasingly significant. Speed harmonization is commonly used as one of the active traffic management strategies. Whether speed harmonization can relieve traffic congestion and improve traffic efficiency becomes a research focus in recent years. Based on the existing speed harmonization models and the traffic characteristics near off-ramp bottleneck, a lane-based speed harmonization model, which searching ways to relieve traffic congestion near off-ramp bottleneck from the main line of urban expressway, was proposed in this paper.

   Firstly, this paper took the recurring bottleneck of Gungong off-ramp as an example. A lane-based transformed cumulative curve of speed was used to identify breakdown on each lane. This method can distinguish the semi-congested state where some lanes are congested while others are not, and deeply analyze the longitudinal and transverse propagation characters of traffic congestion.    Then, the changes of traffic flow rate before and after the traffic congestion in each lane and the flow-velocity diagrams in inside and outside lane of an urban expressway bottleneck near off-ramp were analyzed. It showed that there are large differences in traffic flow characteristics between inside and outside lanes of main line near bottleneck of Gungong off-ramp. Therefore, delicacy management based on each lane is of great significant.

   Then, based on existing cross-sectional speed harmonization model, a lane-based speed harmonization model was designed after modification on in-ramp and off-ramp which can calculate optimal speed on each lane. The METANET traffic flow model was used to describe and forecast the traffic flow status, and shortest travel time as well as maximum total throughput was the optimization objective in this proposed model.

   Finally, based on the VISSIM traffic simulation software, MATLAB software and Visual Basic software, this paper developed simulation system on the case of Gungong off-ramp bottleneck to scientifically evaluate the effectiveness of lane-based and section-based speed harmonization model in different traffic demand situations. The simulation result indicated that, speed harmonization strategy is more adaptive for medium and high traffic flow. When traffic demand is high, both of lane-based model and section-based model can reduce the vehicle average delay and the number of stops compared to no control strategies, which has certain benefits to relieve traffic congestion and improve traffic safety. And in most cases, the laned-based speed harmonization strategy is superior to the section-based model in some respects such as the vehicle average delay, the space-time range of congestion and the number of stops. Therefore, the implementing form of lane-based speed harmonization strategy was further discussed in brief.