A survey of unmanned aerial vehicle flight data anomaly detection: Technologies, applications, and future directions

Flight data anomaly detection plays an imperative role in the safety and maintenance of unmanned aerial vehicles (UAVs). It has attracted extensive attention from researchers. However, the problems related to the difficulty in obtaining abnormal data, low model accuracy, and high calculation cost have led to severe challenges with respect to its practical applications. Hence, in this study, firstly, several UAV flight data simulation softwares are presented based on a brief presentation of the basic concepts of anomalies, the contents of UAV flight data, and the public datasets for flight data anomaly detection. Then, anomaly detection technologies for UAV flight data are comprehensively reviewed, including knowledge-based, model-based, and data-driven methods. Next, UAV flight data anomaly detection applications are briefly described and analyzed. Finally, the future trends and directions of UAV flight data anomaly detection are summarized and prospected, which aims to provide references for the following research.