In this paper, we present a deep hybrid model to detect abnormal flights. Deep hybrid models for anomaly detection use deep neural networks mainly autoencoder as feature extractors, the features learned within the hidden representations of autoencoder are then input to cluster algorithm to detect abnormal flights. The model can detect flight anomalies and associated risks without requiring predefined criteria or domain knowledge. In this paper, 2018 annual flight data of Daocheng Yading airport was taken as the experimental data which contains 981 flights. Our model performed well and 91 abnormal flights were detected.
Flight Anomaly Detection Based on Deep Hybrid Model
14.10.2020
413030 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Multiclass Anomaly Detection in Flight Data Using Semi-Supervised Explainable Deep Learning Model
AIAA | 2022
|Active Learning in Flight Anomaly Detection
AIAA | 2024
|