The invention discloses a track irregularity estimation method based on a CNN-GRU hybrid network. The track irregularity estimation method comprises the steps that firstly, a train body acceleration sequence and a track irregularity sequence detected by a track detection vehicle on a certain railway line are obtained; carrying out abnormal value removal by utilizing a Pauta criterion, filling incomplete data by adopting a sequence average value, and constructing a data set; carrying out normalization processing on the train body acceleration data and the track irregularity data of which the abnormal values are removed, and dividing a training set and a test set; building a CNN-GRU hybrid estimation network, and realizing track irregularity estimation through a GRU network; and estimating a test sample by using the trained network to obtain an estimation result of the track irregularity. According to the method, the CNN is combined with the GRU, and according to the data characteristics and the advantages of the network during the processing of the time sequence, the two basic network models are subjected to advantage complementation by adopting a serial chain structure, so that the track irregularity is estimated.

    本发明公开了一种基于CNN‑GRU混合网络轨道不平顺估计方法,首先获取某段铁路线路上轨道检测车检测的列车车体加速度序列和轨道不平顺序列;利用拉依达准则进行异常值去除,并采用序列平均值填充残缺数据,构建数据集;对去除异常值的列车车体加速度数据和轨道不平顺数据进行归一化处理,并划分训练集与测试集;搭建CNN‑GRU混合估计网络,通过GRU网络实现轨道不平顺估计;利用训练好的网络对测试样本进行估计,得到轨道不平顺的估计结果。本发明通过将卷积神经网络CNN与门控循环单元GRU相结合,根据数据特点以及网络在处理时序序列时自身的优势,采用串联的链式结构将两个基本网络模型进行优势互补,对轨道不平顺进行估计。


    Access

    Download


    Export, share and cite



    Title :

    CNN-GRU-based hybrid network track irregularity estimation method


    Additional title:

    基于CNN-GRU混合网络轨道不平顺估计方法


    Contributors:
    KIM YOUNG-TAEK (author) / LI JIAJIE (author) / LIANG LILI (author) / JI WENJIANG (author) / YI YINGMIN (author) / HEI XINHONG (author) / MU LINGXIA (author) / FEI RONG (author)

    Publication date :

    2023-11-10


    Type of media :

    Patent


    Type of material :

    Electronic Resource


    Language :

    Chinese


    Classification :

    IPC:    B61L Leiten des Eisenbahnverkehrs , GUIDING RAILWAY TRAFFIC / B61D Einzelheiten oder Arten des Wagenaufbaus von Eisenbahnfahrzeugen , BODY DETAILS OR KINDS OF RAILWAY VEHICLES / G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen



    Estimation of Vertical Track Irregularity Based on NARX Neural Network

    Jiang, Jie / Wang, Fuge / Zhang, Yong et al. | British Library Conference Proceedings | 2016


    TRACK IRREGULARITY MEASURING DEVICE

    NAGANUMA YASUKUNI / ARAI TOMOYA / NISHIO TOMOAKI | European Patent Office | 2020

    Free access

    Estimation of Vertical Track Irregularity Based on Extended Kalman Filter

    Wang, Gui / Xing, Zongyi / Wang, Xiaohao et al. | Springer Verlag | 2016


    Estimation of Vertical Track Irregularity Based on Extended Kalman Filter

    Wang, Gui / Xing, Zongyi / Wang, Xiaohao et al. | British Library Conference Proceedings | 2016


    Prediction Method of Track Irregularity Based on EMD

    Li, Zaiwei / He, Yuelei / Shi, Jin et al. | ASCE | 2020