摘 要：行程时间信息在交通管控中起着重要作用。车辆轨迹信息可以提供大规模路网行程时间数据。然而，由于轨迹数据的稀疏性，路网行程时间常出现数据缺失问题。为解决这一问题，构建一种生成对抗网络（GAN）模型。该模型通过拟合数据丰富路段的行程时间数据的概率分布，为数据缺失路段生成行程时间数据。利用滴滴出行轨迹数据进行了数值实验，结果表明GAN 模型的数据补全能力优于对比方法。
关键词：生成对抗网络；行程时间；数据补全
DOI:10.19850/j.cnki.2096-4706.2022.21.021
基金项目：国家自然科学基金资助项目（62002101）
中图分类号：TP311；U491 文献标识码：A 文章编号：2096-4706（2022）21-0088-03
Travel Time Data Imputation for Urban Road Network Based on Generative Adversarial Network
WANG Yiming1, ZHANG Junshuo1, LI Yang2, JING Rongrong1, ZHANG Kunpeng1,3
(1.College of Electrical Engineering, Henan University of Technology, Zhengzhou 450001, China; 2.State Grid Jiaozuo Power Supply Company, Jiaozuo 454000, China; 3.Department of Automation, Tsinghua University, Beijing 100084, China)
Abstract: Travel time information plays an important role in traffic control. Trajectory data from vehicles can provide largescale travel time data of road network. However, due to the sparsity of trajectory data, the problem of missing data often occurs in road network travel time. In order to solve this problem, a Generative Adversarial Network (GAN) model is proposed. By fitting the probability distribution of travel time data from data rich links, the model generates travel time data for data missing links. The numerical experiments are carried out by using trajectory data from DiDi ChuXing, and the results show that the data imputation capability of GAN model is better than that of comparison methods.
Keywords: Generative Adversarial Network; travel time; data imputation
参考文献：
[1] 张腾，林贵敏，邱立达，等 . 基于机器学习的交通数据分析与应用 [J]. 现代信息科技，2018，2（12）：16-18.
[2] LI Y，LI Z，LI L. Missing traffic data：comparison of imputation methods [J].IET Intelligent Transport Systems，2014，8（1）：51-57.
[3] LI L，SU X，ZHANG Y，et al. Trend modeling for traffic time series analysis：An integrated study [J].IEEE Transactions on Intelligent Transportation Systems，2015，16（6）：3430-3439.
[4] LIN Y，DAI X，LI L，et al. Pattern sensitive prediction of traffic flow based on generative adversarial framework [J].IEEE Transactions on Intelligent Transportation Systems，2018，20（6）：2395-2400.
[5] LAÑA I，OLABARRIETA I I，VÉLEZ M，et al. On the imputation of missing data for road traffic forecasting：New insights and novel techniques [J].Transportation research part C：emerging technologies，2018，90：18-33.
[6] DUAN Y，LV Y，LIU Y L，et al. An efficient realization of deep learning for traffic data imputation [J].Transportation research part C：emerging technologies，2016，72：168-181.
[7] ZHUANG Y，KE R，WANG Y. Innovative method for traffic data imputation based on convolutional neural network [J].IET Intelligent Transport Systems，2019，13（4）：605-613.
[8] LIANG Y，CUI Z，TIAN Y，et al. A deep generative adversarial architecture for network-wide spatial-temporal traffic-state estimation [J]. Transportation Research Record，2018，2672（45）：87-105.
[9] CHEN Y，LV Y，WANG F Y. Traffic flow imputation using parallel data and generative adversarial networks [J].IEEE Transactions on Intelligent Transportation Systems，2019，21（4）：1624-1630.
作者简介：王怡茗（2001—），男，汉族，河南驻马店人，本科在读，主要研究方向：交通预测；通讯作者：张坤鹏（1987—），男，河南周口人，讲师，硕导，博士后，主要研究方向：智能交通系统。