基于CEEMD-GRU组合模型的快速路短时交通流预测

沈富鑫 邴其春 张伟健 胡嫣然 高鹏

摘 要：為了提高短时交通流预测精度，提出了基于互补集成经验模态分解（complementary ensemble empirical mode decomposition，CEEMD）和门控循环单元（Gated Recurrent Unit，GRU）组合模型的快速路短时交通流预测方法。首先，运用互补集成经验模态分解算法，将非稳定的原始交通流时间序列数据分解为相对平稳的多个模态分量;然后，将分解后的模态分量分别建立GRU模型进行单步预测;最后，叠加每个分量的预测值，获取最终预测结果，并采用上海市南北高架快速路实测交通流数据进行实例验证。结果表明：CEEMD-GRU组合模型的预测效果明显优于GRU神经网络模型、EMD-GRU组合模型以及EEMD-GRU组合模型，平均预测精度分别提升了33.4%，25.6%和18.3%。CEEMD-GRU组合模型能够有效提取交通流数据特征分量，提高预测精度，为交通管控提供科学决策依据。

关键词：公路运输管理;城市快速路;短时交通流预测;互补集成经验模态分解;门控循环单元神经网络

中图分类号：U491   文献标识码：A

doi：10.7535/hbkd.2021yx05003

收稿日期：2021-06-15;修回日期：2021-08-15;责任编辑：张士莹

基金项目：山东省重点研发计划项目（2019GGX101038）;国家自然科学基金（51678320）;山东省自然科学基金（ZR2019MG012）

第一作者简介：沈富鑫（1992—），男，山东安丘人，硕士研究生，主要从事智能交通系统关键理论与技术方面的研究。

通讯作者：邴其春副教授。E-mail：bingqichun@163.com

Short-term traffic flow prediction of expressway based on CEEMD-GRU combination model

SHEN Fuxin1，BING Qichun1，ZHANG Weijian1，HU Yanran1，GAO Peng2

（1.School of Mechanical and Automotive Engineering，Qingdao University of Technology，Qingdao，Shandong 266520，China;2.Qingdao Transportation Public Service Center，Qingdao，Shandong 266100，China）

Abstract：In order to improve the accuracy of short-term traffic flow prediction，a short-term traffic flow prediction method of expressway based on the combined model of complementary ensemble empirical mode decomposition （CEEMD） and gated recurrent unit （GRU） was proposed.Firstly，the unstable original traffic flow time series data were decomposed into relatively stable multiple modal components by complementary ensemble empirical mode decomposition algorithm.Then，a GRU model was established for each decomposed modal component sequence for one-step prediction.Finally，the predicted value of each component was superimposed to obtain the final prediction result，and the measured traffic flow data of north-south elevated expressway in Shanghai was used to verify and analyze the model.The experimental results show that the prediction effect of CEEMD-GRU combination model is superior to GRU neural network model，EMD-GRU combination model and EEMD-GRU combination model，and the average prediction accuracy is improved by 33.4%，25.6% and 18.3%，respectively.CEEMD-GRU combination model can effectively extract the characteristic components of traffic flow data and improve the prediction accuracy，which provides scientific decision-making basis for traffic control management.

Keywords：

road transportation management;urban expressway;short-term traffic flow prediction;complementary ensemble empirical mode decomposition;gated recurrent unit neural network

精准的短时交通流预测信息是智能交通系统中的关键部分。随着汽车保有量的增加以及人们对通行的要求不断提高，交通污染、交通拥挤、交通事故等一系列交通问题严重影响着城市居民的日常生活品质。……