Fault classification for quadrotor UAV using spatial displacement statistical features

Spatial displacement statistical features are crucial parameters extracted from optical sensor data carrying informative imprints related to the UAV flying conditions. Incorporating these features into a flight faulty classification model for UAVs can provide insights into spatial variations to detect abnormalities and potential faults. This work integrates spatial displacement statistical analysis with a machine learning classification approach to propose a malfunction detection model for vertical take-off and landing quadrotor WAV. Spatial displacement data of the hovering quadrotor in healthy and faulty conditions are recorded using an optical camera system. Pertinent spatial parameters are retrieved using statistical analysis to provide informative features for the classification model. These spatial statistical features indicate hidden flying characteristics concealed by the raw signals captured during flight. The extracted features are used to train and classify different types of faults. The classifier model is also tested to validate its effectiveness in classifying faults from quadrotor UAVs with lower battery levels. The approach shows promising potential for real-time fault diagnosis using spatial displacement sensors to enhance the operational safety of the quadrotor UAV.