Scholarlink

The inertial measurement unit (IMU) is an essential uncrewed aerial vehicle (UAV) sensor among the most safety-critical UAV parts. The drone crashes a few seconds after the IMU issue. Therefore, an efficient fault detection and diagnosis system is essential to ensure flight safety and the successful completion of drone missions. This study proposes a random forest (RF) method with principal component analysis as a feature extraction technique to enhance the accuracy of the used model. The proposed method detects and diagnoses the most frequent blade faults, including fractures and distortions, through the gyroscope and accelerometer dataset. With experimental data, our proposed method has been evaluated and can effectively detect and diagnose thrust faults with a perfect accuracy of \(99\%\) .

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Random Forest Multiclass Classification Fault Detection and Diagnosis for Quadcopter UAVs

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Zineb Adaika,
  • Mohamed Boumehraz,
  • Alaa Abdulhady Jaber

摘要

The inertial measurement unit (IMU) is an essential uncrewed aerial vehicle (UAV) sensor among the most safety-critical UAV parts. The drone crashes a few seconds after the IMU issue. Therefore, an efficient fault detection and diagnosis system is essential to ensure flight safety and the successful completion of drone missions. This study proposes a random forest (RF) method with principal component analysis as a feature extraction technique to enhance the accuracy of the used model. The proposed method detects and diagnoses the most frequent blade faults, including fractures and distortions, through the gyroscope and accelerometer dataset. With experimental data, our proposed method has been evaluated and can effectively detect and diagnose thrust faults with a perfect accuracy of \(99\%\) .