Enhancing piloting techniques in wind-dynamic final approaches using self-organising maps for flight data monitoring

Abstract

Flight data monitoring (FDM) has been implemented in commercial aviation to flag irregular flight parameters for further analysis. However, analysis was limited to exceedance of acceptable range for individual parameters, hindering an integrated evaluation of potential underlying flight safety issues. For instance, pilot’s response towards dynamical wind during approach is critical to ensure landing safety, especially at challenging aerodromes such as Tenerife South Airport, but individual parameters could not reveal such complex scenarios. Hence, this paper proposes a framework by leveraging Quick Access Recorder (QAR) data from FDM with self-organising maps to reveal the pilot interventions in response to wind changes during final approach. Upon clustering, we selected three main clusters from the clustering results and further analysed the pilot’s response and reveal their intervention techniques. Our findings revealed that a mixed use of the autothrottle and the flight director is relevant to the magnitude of the wind changes and yield a more stable approach. Hence, maintaining pitch stability and adjusting the thrust to compensate for the wind changes appear to counteract the wind changes effectively.