Aircraft Flight Movement Anomaly Detection using Automatic Dependent Surveillance Broadcast

Anomaly detection in this research aims to reduce and prevent flight accidents by analyzing abnormal data on aircraft flights using Automatic Dependent Surveillance-Broadcast (ADS-B). ADS-B is an aircraft backup radar that transmits aircraft sensor information via radio frequency. The data may be used to detect significant changes in aircraft by utilizing Deep Learning (DL), where the results potentially serve as study materials for aircraft technicians in conducting maintenance. Bidirectional LSTM (Bi-LSTM) and bidirectional GRU (Bi-GRU) models were proposed using data mining methods. The stages of data mining carried out adopted techniques such as business understanding, data acquisition and understanding, data preparation, modeling, evaluation and performance measurement. Evaluation results indicate that Bi-LSTM achieved better overall accuracy of 99.44% and f1-score of 99.51%. The Bi-LSTM model in this study has the potential to be applied in aircraft flight movement anomaly detection using ADS-B device data.

Keywords: Aircraft flight movement, anomaly detection, aircraft ADS-B device, flight anomalies, data mining.