Assessment of the Detection of Outliers in Rotorcraft Maneuvers Using Neural Network Models

The increasing complexity and diversity of helicopter operations requires advanced safety mechanisms to identify and mitigate potential risks. An important element of helicopter flight is the helicopter maneuvers, with several different possibilities and scenarios, so establishing consistent safety metrics is challenging. Traditional methods, which often rely on fixed safety thresholds, could be inadequate for addressing the dynamic, multi-parameter, unstable, and multi-faceted aspects of helicopter flight. The primary goal of this research is to improve rotorcraft aviation safety by providing a more adaptive and precise approach to detecting operational anomalies experienced during flight maneuvers. This study employs neural networks to enhance the reliability and accuracy of helicopter flight modeling and outlier detection. A comparative assessment of several neural network architectures and outlier detection methods is conducted to evaluate their performance and identify the most effective alternatives for this type of data. The outcomes of this work aim to pave the way for more robust and adaptive safety monitoring systems, with the hope to reduce the incidence of helicopter-related incidents and improve overall operational safety.