基于多维综合交通图谱的高速公路拥堵蔓延消散预测模型研究

Prediction Model of Expressway Congestion Spread and Dissipation based on Multi-dimensional Comprehensive Traffic Patterns

随着经济的高速发展,我国高速公路交通量急剧增加,交通出行在时间和空间上的分布不均导致高速公路交通拥堵频发。对于道路管理者而言,简单的新建道路或道路升级改造无法解决高速公路拥堵的难题,需要着眼于对交通出行的合理引导。分析高速公路交通拥堵影响的时空规律,有利于科学引导公众交通出行,进而提升交通系统运行效率,满足交通管理者管理优化需求。因此,提出一种基于多维综合交通图谱的拥堵蔓延消散预测模型。模型通过聚类道路历史交通量数据生成交通图谱,进而识别预测路段的动态交通量,最终基于预测交通量,建立拥堵蔓延消散模型,实现对拥堵蔓延消散过程的定量评估。以北京市京开高速公路玉泉营桥至菜户营南路北向南方向高速公路路段为例,验证拥堵蔓延消散模型的有效性。结果表明:将路段的动态交通量预测结果作为交通波模型的历史交通量输入,可有效预测路段拥堵蔓延传播速度,模型平均预测准确性可达88.04%。该模型可有效应用于交通诱导等交通管制措施,促进道路交通流合理分配。

With the rapid development of national economy, the traffic volume of expressways increases dramatically in China. The uneven distribution of travel volume in time and space leads to frequent traffic congestion on expressways. For road management authority, simply building new roads or upgrading roads cannot solve such problem, and it is necessary to focus on the reasonable guidance of traffic travel. Analysis of the temporal and spatial law of the impact of highway traffic congestion is helpful to scientific guidance of people’s travel, thus improving the efficiency of transportation system and meeting the needs of stake holders for management optimization. Therefore, in this paper a congestion spread and dissipation forecasting model is proposed according to the comprehensive traffic patterns. By clustering road history traffic flow data, the traffic patterns are recognized, and then forecast is conducted to identify dynamic traffic route. Finally based on the forecast traffic volume, congestion spread and dissipation model is developed and implement quantitative evaluation of the spread of congestion dissipation process. In this paper, the validity of the congestion spread and dissipation model is verified relying on the section of the Beijing-Kaifeng expressway from Yuquanying Bridge to Caihuying South Road. The results show that the prediction results of dynamic traffic volume can be used as the historical traffic volume of the traffic wave model to predict the spread and propagation speed of traffic congestion. The average prediction accuracy of the model can reach 88.04%. The model can be effectively applied to traffic control measures such as traffic guidance to promote rational distribution of road traffic flow.