Machine Learning Based Techniques for Failure Detection and Prediction in Unmanned Aerial Vehicle
Küçük Resim Yok
Tarih
2024
Yazarlar
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
Institute of Electrical and Electronics Engineers Inc.
Erişim Hakkı
info:eu-repo/semantics/closedAccess
Özet
Unmanned aerial vehicles (UAVs) are aircraft with-out human pilot on the board. UAVs have two flight mode: Auto and Manual. In Auto mode, UAV follows a predefined path. The path is embedded in the control system of UAV. In manual mode, human operator in ground control station controls the trajectory of the vehicle remotely. UAVs have diverse applications in military as well as civil sectors. UAVs have to operate in a variety of unseen environments. The diverse usage and uncertainty in operational environment demand safe and reliable operation. Timely identification and rectification of faults stand as a crucial requirement for the operation. One of the major cause of UAV failure is engine fault. In this paper we investigate affectiveness of machine learning techniques regarding engine fault detection and prediction. We analyzed the techniques on AirLab Failure and Anomaly (ALFA) Dataset. For fault detection we used Multi-Layer Perceptron, Random Forest, Support Vector Machine, Ada Boost, Gradient Boosting, Logistic Regression and Single Dimensional Convolutional Neural Network. We observed that Random Forest is most effective technique for fault detection with F-l Score of 0.99. Regarding fault prediction we tried LSTM and GRU based network in different settings. Gated Recurrent Unit performed best with F -1 Score of 0.99 while predicting fault four second ahead of time. © 2024 IEEE.
Açıklama
Anahtar Kelimeler
Failure Detection, Failure Prediction, GRU, Linear Re-gression, LSTM, Machine Learning, Random Forest
Kaynak
2024 IEEE 3rd International Conference on Computing and Machine Intelligence, ICMI 2024 - Proceedings
WoS Q Değeri
Scopus Q Değeri
Cilt
Sayı
Künye
Mustafa, A., Jamil, A., & Hameed, A. A. (2024, April). Machine Learning Based Techniques for Failure Detection and Prediction in Unmanned Aerial Vehicle. In 2024 IEEE 3rd International Conference on Computing and Machine Intelligence (ICMI) (pp. 1-5). IEEE.
