城市主干路交通拥堵预测方法研究

发布时间：2018-11-18 20:46

【摘要】：城市的快速发展吸引了越来越多的交通流,由此带来的交通拥堵问题却成为制约城市发展的一个瓶颈,这是许多城市在面临和亟待解决的问题。各地一方面通过增加道路基础设施建设来缓解交通问题,另一方面结合智能交通系统共同治理拥堵,而结合智能交通系统治理交通拥堵已成为抓缓堵、保畅通、惠民生的重要方法和发展趋势。交通拥堵预测作为智能交通系统的重要组成部分,是缓解城市交通拥堵的有效措施。论文在系统分析交通拥堵概念、分类及量化标准的基础上,针对目前基于交通流参数对城市道路进行交通状态预测的不足,选取了交通流、驾驶员感受及微观驾驶行为三个方面的指标,运用自回归移动平均法(ARIMA)、二次指数平滑法、BP神经网络分别对选取的指标进行预测;基于多模型融合技术建立了组合预测模型,并对各指标进行预测,检验其预测精度,验证了综合预测方法的可行性及有效性。基于城市主干路样本数据,运用层次分析法及K—均值聚类分析法建立了交通状态综合判别算法,将实际的交通状态分类,标记为顺畅、正常、轻度拥挤、重度拥堵四个等级。并用预测数据进行交通状态判别,验证了判别方法的合理性及可靠性。针对日益严重的城市交通拥堵,从宏观策略上及基于预测的交通状态探讨了缓解城市交通拥堵问题的思路。论文尝试从交通流,驾驶员感受及微观驾驶行为三方面的参数建立城市道路交通拥堵综合预测及判别方法,对研究城市道路交通拥堵预测具有一定借鉴意义。
[Abstract]:The rapid development of the city attracts more and more traffic flow, but the traffic congestion becomes a bottleneck that restricts the development of the city, which is the problem that many cities are facing and need to be solved urgently. On the one hand, by increasing the construction of road infrastructure to alleviate traffic problems, on the other hand, by combining with the Intelligent Transportation system (its), the localities have jointly managed congestion, while the combination of the Intelligent Transportation system (its) has become a way to slow down traffic congestion and keep it unblocked. The important method and development trend of benefiting the people. Traffic congestion prediction, as an important part of intelligent transportation system, is an effective measure to alleviate urban traffic congestion. Based on the systematic analysis of the concept, classification and quantification of traffic congestion, the paper selects traffic flow, aiming at the deficiency of traffic state prediction based on traffic flow parameters. Using the autoregressive moving average (ARIMA),) quadratic exponential smoothing method and BP neural network to predict the selected indexes. The combined prediction model is established based on multi-model fusion technique. The prediction accuracy of each index is tested and the feasibility and effectiveness of the comprehensive prediction method are verified. Based on the sample data of urban trunk roads, a comprehensive discrimination algorithm of traffic state is established by using AHP and K-means clustering analysis. The actual traffic states are classified as smooth, normal, mildly congested, and severely congested. The rationality and reliability of the discriminant method are verified by using the prediction data to distinguish the traffic state. In view of the increasingly serious urban traffic congestion, this paper discusses the ways to alleviate the problem of urban traffic congestion from the macro strategy and the traffic state based on forecast. This paper attempts to establish a comprehensive prediction and discrimination method of urban road traffic congestion from three aspects of traffic flow, driver's feeling and micro driving behavior, which has certain reference significance for studying urban road traffic congestion prediction.
【学位授予单位】：长安大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U491.265

【相似文献】