蔡青

      随着城市的发展，城市交通拥堵问题也日益严重，逐渐成为制约我国城市发展的关键因素。对于城市道路交通，交叉口是城市道路的关键节点，是城市交通管理与控制的重要环节，交叉口管理的效果直接影响城市道路的运行。排队长度是交叉口的重要指标，其反映了交叉口排队情况，是交叉口的控制影响因素之一。随着交通检测技术的发展，检测数据越来越多源化，通过多源数据获取排队长度是非常有研究意义的。

      本文的研究主要建立具有普遍适用性的排队长度计算模型对其应用开展研究，主要内容包括：

      1、论文对交叉口排队过程进行了深入分析。详细分析信控交叉口排队过程，根据不同的交通流状态将周期根据排队过程划分为不同的状态，同时将交叉口划分为不同的区域。

      2、论文对交通检测数据进行了分析研究。深入分析了实际的交通检测数据特征，结合交叉口排队过程，提出了利用检测数据判定排队状态的方法，且梳理了排队长度与检测器的不同位置关系。

      3、论文对排队长度计算模型进行了研究。利用多源检测数据，提出了确定排队长度与检测器的不同位置关系的方法。以波动理论为理论基础，根据不同的位置关系，建立了信控交叉口排队长度计算模型。该模型降低了已有计算模型中对数据的要求，并具有普遍的适用性。

      最后，论文结合实际数据进行了案例研究，并分析了模型适用性的影响因素，分析结果验证了本文提出的计算方法具有可行性，说明本文的研究具有十分重要的理论意义和实践价值。

关键词：多源检测数据，信控交叉口，排队长度，波动理论

    With the development of urbanization, traffic congestion has become a serious problem. In the fact, the management and control at intersection are definitely important as the intersection is the key node for the transportation. Queue length which is one of the crucial measurements of signalized intersection which influences the traffic control strategy. With the explosion of traffic sensing data, it is the time to work on building queue length estimation model based on multisource traffic data.

    This paper aims to propose queue length models which can be utilized at different situation. The contents are as follows:

    1.Queuing process analysis. This paper analyzes the process of queuing at signalized intersection, based on which different status and sections are determined.

    2.Traffic data analysis. This paper mines the traffic detecting data, based on which methods of determining queuing status are developed. Besides, four cases are divided based on the space relationship among the rear of initial queue, maximum queue, point sensor location and floating car stop or start-up point.

    3.Queue length estimation models. By using multisource traffic data, this paper develops queue length estimation models based on shockwave theory

    Finally, the models are applied to ground truth data observed from the field to verify its feasibility and practicability.

Key Words:  multisource traffic data, signalized intersection, queue length, shockwave theory