摘要
为提升含缺失值城市道路网络交通流预测的准确性,对生成对抗网络的生成器和判别器进行重构,改进了损失函数,提出了交通流缺失数据补全的交通生成对抗插补网络。基于经验模态分解、图卷积网络和门控循环单元设计了EMD-GCN-GRU城市道路网络交通流预测模型。首先,对交通流数据进行经验模态分解,重构了各同级分量为后续预测模型的输入。然后,利用图卷积网络学习路网拓扑结构,捕获了交通流空间特征,再利用门控循环单元捕获交通流时间特征。最后,利用TGAIN补全含缺失值的路网交通流数据后,再利用EMD-GCN-GRU模型进行交通流预测。采用深圳市罗湖区平均车速数据集,构造了多种具有不同缺失模式和不同缺失比率的交通流数据用于模拟实际缺失情形,在ModelArts开发平台进行方法的有效性验证。结果表明:相较于常用的矩阵分解插补方法,TGAIN模型在数据集随机缺失模式下补全准确性较高,在非随机缺失率低于50%时补全性能较好;与其他预测算法相比,EMD-GCN-GRU交通流预测模型具有更高的预测精度;将数据补全方法TGAIN和交通流预测方法EMD-GCN-GRU相结合进行含缺失值城市道路网络交通流预测,显著降低了数据缺失和数据噪声对交通流预测的负面影响,捕获了路网交通流的时空相关性,进而提升了城市道路网络交通流预测的精度。
In order to improve the accuracy of urban road network traffic flow prediction with missing values,the generator and discriminator of the generative adversarial network are reconstructed,the loss function is improved,and the traffic generative adversarial imputation network(TGAIN)is proposed for the completion of the missing data of traffic flow.Based on empirical mode decomposition(EMD),graph convolutional networks(GCN)and gated recurrent unit(GRU),EMD-GCN-GRU model is designed for urban road network traffic flow prediction.First,the traffic flow data is processed by empirical mode decomposition and each component of the same level is reconstructed as the input of the subsequent prediction model.Then,the graph convolutional networks are used to learn the road network topology to capture the spatial characteristic of traffic flow,and the gated recurrent unit is employed to capture the temporal characteristic of traffic flow.Finally,for the road network traffic flow data with missing values,TGAIN is used to complete the data,and then EMD-GCN-GRU is used to predict the traffic flow.The average vehicle speed data set of Luohu districtin in Shenzhen is used to construct a variety of typical traffic flow data with different missing patterns and different missing rates to simulate the actual missing situation.The effectiveness of the method is verified on the ModelArts development platform.The result shows that(1)compared with the commonly used matrix factorization imputation method,the TGAIN model has higher completion accuracy in the random missing mode of the dataset and has better completion performance when the non-random missing rate is lower than 50%;(2)compared with other prediction algorithms,the proposed prediction method has higher prediction accuracy;(3)combining the data imputation method TGAIN with the traffic flow prediction method EMDGCN-GRU for urban road network traffic flow prediction with missing values can significantly reduce the negative impact of missing data and data noise on traffic flow prediction and capture the spatial and temporal correlation of network traffic flow,which improves the accuracy of urban road network traffic flow prediction.
作者
陈建忠
吕泽凯
蔺皓萌
CHEN Jian-zhong;LÜZe-kai;LIN Hao-meng(School of Automation,Northwestern Polytechnical University,Xi’an Shaanxi 710129,China)
出处
《公路交通科技》
CSCD
北大核心
2023年第9期205-214,共10页
Journal of Highway and Transportation Research and Development
基金
国家自然科学基金项目(11772264)
陕西省自然科学基础研究计划项目(2020JM-119)。
关键词
智能交通
交通流预测
生成对抗插补网络
城市道路网络
图卷积网络
ITS
traffic flow prediction
generative adversarial imputation network
urban road network
graph convolutional network