In order to improve the hitting accuracy of inertial-guided launch vehicle, the outliers in the output Euler angles and accelerations of Inertial Navigation System (INS) must be detected before launching. Current outlier detection methods in INS are mainly supervised approaches and may result in too many false positives. In this paper, we investigate and introduce the Long Short-Term Memory (LSTM) recurrent neural networks to forecast time sequences in INS. Due to its gate mechanism and capability to maintain internal state, our method could get relatively low prediction errors. Once the prediction model is trained, we propose a method to set a threshold to detect the outliers. We introduce the Exponentially Weighted Moving Average (EWMA) method to dampen spikes in prediction errors due to abrupt changes in the outputs of INS. Finally, we carried out experiments on the real outputs of INS using our method to detect outliers. The results demonstrate the suitability of LSTM networks for outlier detection.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Detecting Inertial Navigation System Outliers Using Long Short-Term Memory Recurrent Neural Networks


    Additional title:

    Lect. Notes Electrical Eng.


    Contributors:
    Yan, Liang (editor) / Duan, Haibin (editor) / Yu, Xiang (editor) / Xiang, Gang (author) / Zhou, Jianming (author) / Peng, Yu (author) / Liu, Qingzhu (author)


    Publication date :

    2021-10-30


    Size :

    13 pages





    Type of media :

    Article/Chapter (Book)


    Type of material :

    Electronic Resource


    Language :

    English






    Vision-Aided Inertial Navigation with Modeling of Measurement Outliers

    Yang, Chun / Soloviev, Andrey / Veth, Michael et al. | AIAA | 2015


    Real-Time Crash Risk Prediction using Long Short-Term Memory Recurrent Neural Network

    Yuan, Jinghui / Abdel-Aty, Mohamed / Gong, Yaobang et al. | Transportation Research Record | 2019


    Forecasting dynamic public transport Origin-Destination matrices with long-Short term Memory recurrent neural networks

    Toque, Florian / Come, Etienne / El Mahrsi, Mohamed Khalil et al. | IEEE | 2016