A KNNS based anomaly detection method applied for UAV flight data stream

Flight data is a kind of time correlated data collected by the aircraft on-board sensor system. Anomaly detection for the flight data turns out to be one of the most essential concern in the area of UAV flight control. In this research, a real time anomaly detection method for the UAV flight data is developed. In this method, a KD-Tree is built to index the historical flight data, KNNS (K Nearest Neighbor Search) is performed to each observed data to identify different types of anomaly and provide a reasonable predicted value. Experiment is carried out to prove the method is more efficient and accurate than the traditional real time anomaly detection methods. Furthermore, the research also tries to find out how the parameter settings influence the method performance which provide a useful guideline for the parameter selection.