Research on Helicopter Flight Safety Detection Algorithm Based on Weighted Calculation
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
In recent years, helicopter flight control technology has developed rapidly, and research on helicopter flight safety has become a hot topic in flight data research and analysis both domestically and internationally. This paper addresses the problem of helicopter flight safety monitoring. Using trajectory data and attitude data, which reflect real-time position in helicopter flight training data, as the main data source, a weighted calculation-based instantaneous anomaly detection algorithm is proposed. This algorithm integrates five unsupervised anomaly detection methods: Local Outlier Factor (LOF), K-Nearest Neighbors (KNN), Isolation Forest (Iforest), Principal Component Analysis (PCA), and Lightweight Online Anomaly Detector (LODA). It uses a feedforward neural network to perform weighted integration of the outputs of various algorithms, with AUC as the evaluation metric, to optimize the comprehensive performance of anomaly discrimination. The algorithm is validated on multiple public datasets and real helicopter flight data. The results show that the weighted fusion method is superior to a single algorithm in terms of average AUC value, and can effectively identify abnormal behaviors in flight states, with strong robustness and practicality. The study also extracts altitude, pitch angle, roll angle, and airspeed as key features from actual flight data, and analyzes anomaly patterns under multi-platform and single-platform scenarios, providing data-driven diagnostic methods for helicopter flight safety.
Fields of Study & Subjects
Anomaly Detection Techniques and Applications
Fault Detection and Control Systems
Aerospace and Aviation Technology
Publication details
- Journal / source
- IEEE International Conference on Power, Electronics and Green Energy (ICPEGE)
- Pages
- 1028–1033
- Publication date
- 28 January 2026
- Language
- EN