Condition monitoring based on flight data for unmanned aerial vehicles (UAVs) is critical since it can enhance the efficiency of task execution and guarantee the flight safety of UAVs. However, UAV flight data are high-dimensional, real-time, and complex dynamic, which means it is difficult to predict. Therefore, there is an urgent need for methods with a powerful capability to process time-series data and predict future conditions. To fully extract the feature for flight data of UAVs, this paper, thus, proposes a Transformer-based condition monitoring method which can find anomalies and provide timely warning to adopt corrective. First, flight data are collected and pre-processed, and after correlation analysis, the parameters with high correlation with the monitored parameter are selected for feeding into the Transformer network, so a Transformer-based forecasting model can be built. Then, it is determined whether the data points outside the thresholds defined by the predicted residual obtained from the forecasting model and if so, it will be considered abnormal. The real flight data is finally applied to verify the efficiency and superiority of the proposed scheme. Experimental results demonstrate that Transformer outperforms comparison approaches in terms of detecting the anomalies of flight data from UAVs.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    A Transformer-Based Condition Monitoring Method for UAV


    Additional title:

    Lect. Notes Electrical Eng.


    Contributors:
    Qu, Yi (editor) / Gu, Mancang (editor) / Niu, Yifeng (editor) / Fu, Wenxing (editor) / Su, Zhou (author) / Qi, Boxun (author) / Wang, Benkuan (author) / Liu, Datong (author)

    Conference:

    International Conference on Autonomous Unmanned Systems ; 2023 ; Nanjing, China September 09, 2023 - September 11, 2023



    Publication date :

    2024-04-23


    Size :

    11 pages





    Type of media :

    Article/Chapter (Book)


    Type of material :

    Electronic Resource


    Language :

    English