|本期目录/Table of Contents|

[1]康晋滔,成卫,张灵.基于二元特征的断面交通数据异常检测与修正[J].中国安全生产科学技术,2019,15(9):70-75.[doi:10.11731/j.issn.1673-193x.2019.09.011]
 KANG Jintao,CHENG Wei,ZHANG Ling.Anomaly detection and correction of sectional traffic data based on bivariate features[J].JOURNAL OF SAFETY SCIENCE AND TECHNOLOGY,2019,15(9):70-75.[doi:10.11731/j.issn.1673-193x.2019.09.011]
点击复制

基于二元特征的断面交通数据异常检测与修正
分享到:

《中国安全生产科学技术》[ISSN:1673-193X/CN:11-5335/TB]

卷:
15
期数:
2019年9期
页码:
70-75
栏目:
职业安全卫生管理与技术
出版日期:
2019-09-30

文章信息/Info

Title:
Anomaly detection and correction of sectional traffic data based on bivariate features
文章编号:
1673-193X(2019)-09-0070-06
作者:
康晋滔1成卫1张灵2
(1.昆明理工大学 交通工程学院,云南 昆明 650504;
2.昆明市公安局交警支队科信处,云南 昆明 650000)
Author(s):
KANG Jintao1 CHENG Wei1 ZHANG Ling2
(1.School of Traffic Engineering,Kunming University of Science and Technology,Kunming Yunnan 650504,China;
2.Science and Information Office,Traffic Police Branch of Kunming Public Security Bureau,Kunming Yunnan 650000,China)
关键词:
智能交通异常检测与修正断面交通数据多元高斯灰色马尔可夫模型
Keywords:
intelligent transportation anomaly detection and correction sectional traffic data multivariate Gauss grey Markov model
分类号:
X951;U491
DOI:
10.11731/j.issn.1673-193x.2019.09.011
文献标志码:
A
摘要:
为了确保交通数据的准确性以及为后续研究提供数据质量保障。提取断面交通数据的两元特征向量速度与流量,采用多元高斯模型拟合两元特征概率密度分布,并利用matlab训练模型,对测试数据集进行异常检测;针对检测出的异常数据点,引入小时窗口,建立修正数据集。随后在各修正窗口内采用波动性处理优化的灰色马尔可夫模型,对上述异常点进行修正。以汕昆高速K2077断面为例,对其采集数据进行模拟计算及分析,结果表明:多元高斯模型对断面交通数据的二元特征拟合效果良好,异常检测作用突出;引入修正窗口简化了数据修正的运算过程,避免了大量历史数据冗余使用;波动性处理优化弥补了灰色马尔可夫模型针对随机性数据性能下降的缺陷,大幅提升了数据修正的准确率。
Abstract:
In order to ensure the accuracy of traffic data and provide data quality assurance for the followup research,the bivariate eigenvector velocity and flow of sectional traffic data were extracted.The probability density distribution of bivariate features was fitted by using the multivariate Gauss model,and the anomaly detection of the testing data sets was carried out by using the MATLAB training model.Aiming at the detected abnormal data points,the hourly window was introduced to establish the corrected data set.Then the grey Markov model optimized by volatility processing was applied in each correction window to correct the above abnormal points.Taking the K2077 section of ShanKun expressway as an example,the collected data were simulated and analyzed.The results showed that the fitting effect of multivariate Gauss model on the bivariate features of sectional traffic data was good,with a prominent role in anomaly detection.The introduction of correction window simplified the operation process of data correction,and avoided the redundant use of large amount of historical data.The optimization of volatility processing made up for the shortcoming of grey Markov model in the performance reduction of random data,and greatly improved the accuracy of data correction.

参考文献/References:

[1]刘冬梅,王文静,杨子帆,等.互联网+时代众包交通大数据应用机制研究[J].公路交通科技,2018,35(7):120-127.LIU Dongmei,WANG Wenjing,YANG Zifan.Study on application mechanism of internet+epoch crowdsourcing traffic big data[J].The Journal of Highway and Transportation Research and Development,2018,35(7):120-127.
[2]Texas Transportation Institute.Advances in traffic data collection and management[R].Washington:Federal Highway Administration,2002.
[3]JACOBSON,NIHAN N L,BENDER J D.Detecting erroneous loop detector data in a freeway traffic management system[J].Transportation Research Board,1990:151-160.
[4]ZHONG M,LINGRAS P,SHARMA S.Estimation of missing traffic counts using factor,genetic,neural,and regression techniques[J].Transportation Research Part C,2004,12(2):139-166.
[5]VANAJAKSHI L,RILETT L R.Loop detector data diagnostics based on Conservation-of-Vehicles principle[J].Transportation Research Record,2004 (1870):162-169.
[6]MIN W,WYNTER L.Real-time road traffic prediction with spatio-temporal correlations[J].Transportation Research Part C:Emerging Technologies,2011,19(4):606-616.
[7]姜桂艳,牛世峰,李红伟.动态交通数据质量评价方法研究[J].北京工业大学学报,2011,37(8):1190-1195.JIANG Guiyan,NIU Shifeng,LI Hongwei.Research on quality evaluation method of dynamic traffic data[J].Journal of Beijing University of Technology,2011,37(8):1190-1195.
[8]易良友.道路交通流量数据处理与预测模型研究[D].重庆:重庆大学,2011.
[9]郭敏,蓝金辉,李娟娟,等.基于灰色残差GM(1,N)模型的交通流数据恢复算法[J].交通运输系统工程与信息,2012,12(1):42-47.GUO Min,LAN Jinhui,LI Juanjuan,et al.Traffic flow data recovery algorithm based on gray residual GM (1,N) model[J].Journal of Transportation Systems Engineering and Information Technology,2012,12(1):42-47.
[10]徐程,曲昭伟,陶鹏飞,等.动态交通数据异常值的实时筛选与恢复方法[J].哈尔滨工程大学学报,2016,37(2):211-217.XU Cheng,QU Zhaowei,TAO Pengfei,et al.Methods of real-time screening and reconstruction for dynamic traffic abnormal data[J].Journal of Harbin Engineering University,2016,37(2):211-217.
[11]范光鹏,孙仁诚,邵峰晶.基于LSTM和Kalman滤波的公交车到站时间预测[J].计算机应用与软件,2018,35(4):91-96.FAN Guangpeng,SUN Rencheng,SHAO Fengjing.Bus arrival time prediction based on LSTM and Kalman filtering[J].Computer Applications and Software,2018,35(4):91-96.
[12]SUN B,MA L,CHENG W,et al.An improved k-nearest neighbors method for traffic time series imputation [C]//Chinese Automation Congress (CAC),2017.
[13]KIMT Y,CHO S B.Web traffic anomaly detection using C-LSTM neural networks[J].Expert Systems with Applications,2018,106:66-76.
[14]曹建,施式亮,曹华娟,等.基于GM(1,1)-Markov的危化品道路运输事故与交通事故预测及关系研究[J].中国安全生产科学技术,2019,15(1):26-31.CAO Jian.SHI Shiliang,CAO Huajuan,et al.Study on prediction and relationship of road transportation accidents of dangerous chemicals and traffic accidents based on GM(1,1)-Markov model[J].China Safety Science and Technology,2019,15(1):26-31.
[15]王曦,祝付玲.基于高斯混合分布的交通拥堵评价模型[J].公路交通科技,2011,28(2):127-132.WANG Xi,ZHU Fuling.Evaluation model of traffic congestion based on gaussian mixture distribution[J].The Journal of Highway and Transportation Research and Development,2011,28(2):127-132.
[16]王晓源,吴芳,朴基男.基于粗集理论的交通流丢失数据补齐方法[J].交通运输工程学报,2008,8(5):91-94.WANG Xiaoyuan,WU Fang,PIAO Jinan.Filling method of missing data for traffic flow based on rough set theory[J].Journal of Traffic and Transportation Engineering,2008,8(5):91-94.

相似文献/References:

备注/Memo

备注/Memo:
收稿日期: 2019-08-26
* 基金项目: 国家自然科学基金项目(61364019)
作者简介: 康晋滔,硕士研究生,主要研究方向为交通信息工程及控制。
通信作者: 成卫,博士,教授,主要研究方向为交通控制,智能交通。
更新日期/Last Update: 2019-10-15