考虑降雨影响的城市交通流车速多层线性模型研究

发布时间：2018-06-20 22:24

本文选题：交通工程 + 交通流　；参考：《公路交通科技》2014年10期

【摘要】：应用多层线性模型研究了降雨强度对城市道路交通流平均行驶车速的影响。研究资料包括道路特征数据(25个路段)、广州市城市道路交通流平均行驶车速数据(有效车速样本量超过160万)和降雨量数据(样本量93 462)。采集时间从2011年6月至2012年4月一共330 d。基于以上大规模数据,应用多层线性模型(HLM),建立了一个降雨条件下道路交通流车速的短时动态预测模型。结果表明:交通流信息存在嵌套、分层的数据结构,相较于传统交通流研究方法,多层线性模型对交通流影响因素的研究将更为合理;降雨对车速的影响具有普遍的空间变异性,这是因为每个路段具有不同的物理特征;交通流的影响因素之间存在着显著的跨层交互效应,这与传统理论对交通流的认识不同;本文提出的降雨条件下交通流车速的短时动态预测模型具有路段适用性广的突出优点,这与大多数传统交通流模型适用范围较为狭窄的特点较为不同。
[Abstract]:The effect of rainfall intensity on the average speed of urban road traffic flow is studied by using multilayer linear model. The data include road characteristic data (25 sections), average driving speed data of Guangzhou city road traffic flow (the sample number of effective speed is more than 1.6 million) and rainfall data (93,462). The collection time was 330 days from June 2011 to April 2012. Based on the large scale data mentioned above, a short-term dynamic prediction model of road traffic flow speed under rainfall condition is established by using the multi-layer linear model (HLM). The results show that there are nested and hierarchical data structures in traffic flow information. Compared with the traditional traffic flow research method, the multi-layer linear model will be more reasonable to study the factors affecting traffic flow. The influence of rainfall on speed has a general spatial variability, which is due to the different physical characteristics of each section, the significant cross-layer interaction between the influencing factors of traffic flow, which is different from the traditional theory of the understanding of traffic flow. The short-time dynamic prediction model of traffic flow speed under rainfall has the prominent advantage of wide applicability, which is different from the narrow range of most traditional traffic flow models.
【作者单位】：中山大学广东省智能交通重点实验室;东莞中山大学研究院智能交通工程中心;
【基金】：广东省科技计划项目(2011A011305002,2011B010700092)
【分类号】：U491.112

【参考文献】