The evaluations of the impact of the pilot’s visual behaviours on the landing performance by using eye tracking technology

Abstract

Introduction. Eye tracking technology can be used to characterise a pilot's visual behaviour as well as to further analyse the workload and status of the pilot, which is crucial for tracking and predicting pilot performance and enhancing flight safety. Research questions. This research aims to investigate and identify the visual-related factors that could affect the pilot's landing operation performance (depending on whether the landing was successful or not). Method. There are 23 participants who performed the task of landing in the Future system simulator (FSS) while wearing eye trackers. Their eye tracking parameters including proportion of fixation count on primary flight display (PFC on PFD), proportion of fixation count on out the window (PFC on OTW), percentage change in pupil diameter (PCPD) and blink count were trained for classification using XGBoost according to whether they landed successfully or not. Results & Discussion. The results demonstrated that eye-movement features can be used to classify and predict a pilot's landing performance with an accuracy of 77.02%. PCPD and PFC on PFD are more crucial for performance classification out of the four features. Conclusion. It is practical to classify and predict pilot performance using eye-tracking technologies. The high importance of PCPD and PFC on PFD indicates that there is a correlation between pilots’ workload and attention distribution and performance, which has important implications for future predictive and analytical research on performance. The prediction of performance using eye tracking suggests that pilot status monitoring has a useful application in flight deck design.