快速路是城市交通的主动脉，然而随着我国经济的高速发展，城市化率稳步提升，城市人口逐年增长，快速路已经难以承载日益扩大的交通需求，拥堵越来越严重。本文以正确认识快速路的通行能力、提升快速路管理与控制的前瞻性和预见性、丰富快速路的监管控制手段为突破口，对快速路动态通行能力进行了建模与估计，实现了快速路交通崩溃事件的预测，在智能车路协同背景背景下，提出了一种有效的可变限速控制策略，对快速路进行实时、主动优化控制。 首先，针对传统理论通行能力在描述快速路实时服务能力上的缺陷，本文提出了快速路动态通行能力的一种描述，确定了四类主要影响因素，分析了各因素对动态通行能力的影响机理，并采用四个参数将其进行了量化，使用理论推导、实际数据分析、软件仿真相结合的方式，获取了各参数与动态通行能力的影响关系，建立了快速路动态通行能力的计算模型。此外，针对数据获取困难、不需要精确计算的场景，本文提出了一种基于Dempster-Shafer证据理论的快速路动态通行能力估计算法，并采用实例验证了该方法的有效性。 其次，针对引发快速路通行效率降低的交通崩溃事件，本文引入区分车道的方法，对快速路的交通崩溃事件进行了分级定义，并根据交通崩溃事件的特点，以道路占有率为显状态，道路崩溃状态为隐状态，建立了隐马尔可夫模型，用以描述二者之间的关系，采用Viterbi算法对该模型进行了求解，实现了交通崩溃状态的预测，获得了良好的预测效果。 然后，为了降低可能发生的交通崩溃事件对快速路通行效率造成的影响，本文引入可变限速控制对相关路段进行主动控制。借助智能车路协同系统在信息感知与发布方面的强大功能，通过引入微观交通信息，对传统的METANET模型进行了改造，建立了更为精确的微观METANET模型。在此基础上，设计了可变限速控制策略，辅以ALINEA匝道控制算法，提出了一套可行的可变限速控制方案。 最后，为了验证微观METANET模型、可变限速控制策略、匝道联合控制策略的有效性，本文采用VISSIM仿真软件，以实际道路和交通流数据为基础，搭建了仿真平台。仿真结果表明， 微观化METANET模型具有良好的交通流参数预测效果，可变限速控制、匝道联合控制策略均能有效地改善道路的交通状态。
Urban expressway is the artery of urban traffic network. However, with the rapid development of economy in our country, the urbanization rate is rising steadily and the urban population increases year by year. Thus, the growing traffic demond exceeds the service capabilities of expressways. The purposes of the research are to form a correct understanding of the traffic capacity of expressway, improve the prospective and predictability of traffic management and control, enrich the methods of traffic monitoring and management. In the context of Intelligent Vehicle Infrastructure Cooperative Systems (IVICS), the paper proposes a computational model and an estimate method of dynamic capacity of expressways, predicts the traffic breakdown state and presents a variable speed limit strategy for real time active traffic management and control. First, the paper presents a new description of dynamic traffic capacity of expressway which remedies the defect of traditional theory capacity in real time service capability measurement. Four types of factors which affect the dynamic capacity are identified, and their influence mechanisms are analyzed. Four parameters are used to quantify the factors. The relationships between the parameters and dynamic capacity are obtained by theoretical derivation, analysis of actual data and software simulation. Then, a model is established for dynamic capacity calculation. In addition, an algorithm for rough dynamic capacity estimation based on Dempster-Shafer evidence theory is proposed and the validity of the method is verified. Next, a lane-based method is used to define traffic breakdown which will probably result in lower traffic efficiency. According to the characteristics of traffic breakdown, the paper takes the road occupancy as the observation state, the traffic breakdown state as the hidden state. Then a hidden Markov model is established to describe their relationships. The Viterbi algorithm is introduced to solve the model so that the traffic breakdown prediction is realized. Then, a variable speed limit strategy is proposed based on a microcosmic METANET model to reduce the impact of possible traffic breakdown. Taking advantage of IVICS in information perception, a more accurate microcosmic METANET model, with which the parameters of freeway can be predicted more accurately, is established by reforming the traditional METANET model. Based on the microcosmic METANET model, a variable speed limit strategy is proposed and a coordination ramp metering strategy based on ALINEA algorithm is introduced. Finally, to verify the validity of microcosmic METANET model, the variable speed limit strategy and the combined strategy, a VISSIM based simulation platform is established according to actual expressway segment and traffic flow data. The results show that the microcosmic METANET model works well in traffic flow data prediction and the two control strategies perform well in improving traffic status.
