Downtime reduction is one of the top priorities for commercial vehicles providers. The major reasons for long downtime include vehicle failures in the middle of the road or trips, prolonged service time due to lack of availability of parts and technician. Furthermore vehicle failures in the mid of the road trips pose danger to the nearby passing vehicles and pedestrians. Huge expenses are observed in the delayed repair due to the fact that failed parts can deteriorate other components. In order to prevent the risks of component failures and huge costs, a deep learning based system was implemented to provide predictive warning before the actual failure. A novel method has been proposed to mimic domain expert's abnormality detection process using GAN (Generative Adversarial Network): Generator in the GAN was used to generate expected normal behavior; discriminator was used to distinguish normal and abnormal behaviors. The prediction score of Machine Learning (ML)/Deep Learning (DL) of generated expected normal behavior, was used as a threshold. Real world Isuzu vehicle data was used to validate the complete pipeline, advanced warning capability was implemented, and validation was shown. A complete pipeline of infrastructure and software development were introduced in this paper.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Time Series Anomaly Detection Based on GAN


    Contributors:
    Sun, Yong (author) / Yu, Wenbo (author) / Chen, Yuting (author) / Kadam, Aishwarya (author)


    Publication date :

    2019-10-01


    Size :

    2506862 byte




    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English



    Anomaly Scoring for Prediction-Based Anomaly Detection in Time Series

    Li, Tianyu / Comer, Mary L. / Delp, Edward J. et al. | IEEE | 2020


    MIM-GAN-based Anomaly Detection for Multivariate Time Series Data

    Lu, Shan / Dong, Zhicheng / Cai, Donghong et al. | IEEE | 2023


    Adaptable Anomaly Detection in Traffic Flow Time Series

    Alam, Md Rakibul / Gerostathopoulos, Ilias / Amini, Sasan et al. | IEEE | 2019


    Anomaly detection in time series via robust PCA

    Jin, Yongjun / Qiu, Chenlu / Sun, Lei et al. | IEEE | 2017


    Telemanom: An Extensible Framework for Time-series Anomaly Detection

    Soderstrom, Tom / Mattmann, Chris / Colwell, Ian et al. | NTRS | 2019