In complex substation environments, the flight safety problem when unmanned aerial vehicles perform inspection tasks becomes more and more prominent. Currently, the flight control system is the main means of UAV flight status monitoring, but it lacks in-depth analysis of flight data, especially flight trajectory data. To make full use of the flight control system data, this paper proposes an anomaly detection and trajectory prediction method. Firstly, the local temporal features of the original data in the subspace are extracted, and the anomaly detection of the flight data is achieved by measuring the changes of the data subspace vectors on the basis of reducing the computational complexity of the data, followed by dynamically adjusting the thresholds of the anomalous data by using one-class support vector machines. And finally, the UAV trajectory prediction is performed by using a time-series based LSTM neural network, focusing on the anomalies that may affect the safety of flights Data Early Warning.
Research on UAV Data Anomaly Detection and Early Trajectory Warning Technology
17.01.2025
877268 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Trajectory anomaly detection system and online trajectory anomaly detection method
Europäisches Patentamt | 2024
|Traffic anomaly detection early warning method and system based on artificial intelligence
Europäisches Patentamt | 2024
|Expressway confluence area vehicle conflict early warning method based on trajectory data
Europäisches Patentamt | 2024
|TRAJECTORY DEVIATION EARLY WARNING METHOD, TERMINAL, AND STORAGE MEDIUM
Europäisches Patentamt | 2024
|