施佳呈 吴戈 俄文娟

摘要：为了准确识别和预测区域交叉口的交通状态，研究道路网络中的拥堵传播规律，以交叉口流量数据为研究基础，提出了一种改进型需求容量的交叉口交通状态判别方法。结合交通领域的具体特点，对算法数据库进行适应性调整，采用一种基于路网拓扑结构约束的Aprior算法，建立区域交叉口拥堵传播规律挖掘模型，并以苏州市某区域的交叉口数据为实例进行验证，从时间维度分析发现拥堵具有相似性和反复性，从空间维度研究获取“回溢”和“流出”2种拥堵传播模式，与现实情况相符。结果表明，模型能够有效挖掘区域交叉口拥堵传播规律，并根据传播规律中各个交叉口拥堵发生的时序特性进行交通状态预测。所得结论对实现拥堵预警和交通拥堵的早期干预具有一定的借鉴意义。

关键词：交通运输系统工程;区域交叉口;拥堵传播规律;交通状态判别;Aprior算法;交通预测

中图分类号：U121文献标识码：A

doi： 10.7535/hbgykj.2020yx03008

Abstract：

In order to recognize and predict the traffic status at regional intersections accurately， as well as to research congestion propagation rules in road networks， an improved demandcapacity method was used to identify the traffic status at intersections based on traffic flow data. Firstly， adaptive adjustment to the database was conducted according to the specific features of the traffic field. Then， the model of mining traffic congestion propagation rules at regional intersections was established by the Aprior algorithm based on the topological constraints of the road network. Finally， intersection data from an area in Suzhou was used to verify the proposed model. It was found that the congestion had similar and repetitive features from the view of the time dimension， and the congestion propagation modes of "spillback" and "outflow" could be obtained from the view of space， which were consistent with the present reality. It indicates that the model can effectively analyze congestion propagation rules at regional intersections and predict traffic status according to the timing characteristics of the congestion at each intersection in the propagation law， which is of great significance for the early warning and intervention of the traffic congestion.

Keywords：

traffic and transportation system engineering; regional intersection; congestion propagation rule; traffic status identification; Aprior algorithm; traffic prediction

交通擁堵是现代化城市亟待解决的交通难题，对交通状态的准确预测是进行交通控制和疏导以缓解交通拥堵的基础。交叉口作为道路网络的节点，容易成为交通拥堵的诱发点。

现有对交叉口交通状态预测的研究成果主要有基于统计分析的预测模型、非线性理论模型、基于仿真的预测模型、智能预测模型等[1]。传统的预测方法大多基于单个交叉口的交通流时间变化特性，随着研究的深入，学者们开始从道路网络的整体分析出发，整合交通流时空变化特性，相关研究大致可分为2类。……