High Power Laser and Particle Beams, Volume. 37, Issue 4, 043006(2025)

Construction and evaluation method of unmanned aerial vehicle faults simulation dataset

Yicheng Wang, Mengjuan Chai, Daojie Yu*, Yijie Bai, Liyue Liang, Tao Li, Jiale Zhou, Jianping Du, and Zhenning Yao
Author Affiliations
  • School of Information Systems Engineering, Information Engineering University, Zhengzhou 450002, China
  • show less
    Figures & Tables(10)
    Self-assembled quadrotor unmanned aerial vehicle (UAV) model and system composition
    Four classic fault types
    ROC curves of the three models for different fault types
    P-R curves of the three models for different fault types
    Confusion matrices of the three models
    • Table 1. Selected state parameters of dataset

      View table
      View in Article

      Table 1. Selected state parameters of dataset

      No.variableNo.variable
      1north velocity14pitch velocity
      2east velocity15yaw velocity
      3down velocity16motor 1
      4acceleration x-component17motor 2
      5acceleration y-component18motor 3
      6acceleration z-component19motor 4
      7roll angle20absolute pressure
      8pitch angle21differential pressure
      9yaw angle22pressure altitude
      10x magnetic field (Gaussian)23channel pitch input
      11y magnetic field (Gaussian)24channel roll input
      12z magnetic field (Gaussian)25channel throttle input
      13roll velocity26channel yaw input
    • Table 2. Fault types with their corresponding labels and sample size

      View table
      View in Article

      Table 2. Fault types with their corresponding labels and sample size

      labelfault modesample sizelabelfault modesample size
      C0health340 850C4roll rate bias fault (sensor)284 932
      C140% reduction in efficiency (single actuator)245 180C5roll rate lock fault (sensor)226 528
      C2bias fault (single actuator)242 898C6roll rate scale fault (sensor)258 065
      C340% reduction in efficiency (dual actuator)254 984C7roll rate drift fault (sensor)236 403
    • Table 3. Confusion matrix

      View table
      View in Article

      Table 3. Confusion matrix

      referencepositive predictionnegative prediction
      positivetrue positive (TP)false negative (FN)
      negativefalse positive (FP)true negative (TN)
    • Table 4. Fault diagnosis performance of the three models on the test set

      View table
      View in Article

      Table 4. Fault diagnosis performance of the three models on the test set

      modelaccuracyprecisionrecall
      WDCNN0.769 10.828 10.769 1
      ResNet0.888 40.901 00.888 4
      QCNN0.918 30.928 00.918 3
    • Table 5. Performance of three compared models on each fault type

      View table
      View in Article

      Table 5. Performance of three compared models on each fault type

      fault modeprecisionrecallF1 scoreprecisionrecallF1 scoreprecisionrecallF1 score
      WDCNNResNetQCNN
      C00.605 10.9730.746 20.883 30.4920.632 00.891 80.9810.934 3
      C10.944 00.7920.861 30.946 00.7880.859 80.739 10.9490.831 0
      C20.917 50.0890.162 30.964 70.9830.973 70.977 60.6120.752 8
      C30.954 40.9010.927 00.772 40.9060.833 90.884 30.9020.893 1
      C40.302 80.4370.357 80.685 20.9470.795 10.966 10.9130.938 8
      C51.000 00.9820.990 91.000 01.0001.000 01.000 01.0001.000 0
      C61.000 00.9990.999 51.000 01.0001.000 01.000 01.0001.000 0
      C70.900 70.9800.938 70.95660.9910.973 50.964 90.9890.9768
    Tools

    Get Citation

    Copy Citation Text

    Yicheng Wang, Mengjuan Chai, Daojie Yu, Yijie Bai, Liyue Liang, Tao Li, Jiale Zhou, Jianping Du, Zhenning Yao. Construction and evaluation method of unmanned aerial vehicle faults simulation dataset[J]. High Power Laser and Particle Beams, 2025, 37(4): 043006

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Save article for my favorites
    Paper Information

    Category:

    Received: Sep. 23, 2024

    Accepted: Dec. 11, 2024

    Published Online: May. 15, 2025

    The Author Email: Daojie Yu (yudj2003@163.com)

    DOI:10.11884/HPLPB202537.240340

    Topics