Investigating air traffic controllers' situation awareness and workload using eye-tracking

• Main Author: Wong, Martin Jun Wei
• Other Authors: Chen Chun-Hsien
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2024
• Subjects: Other; Situation awareness; Workload; Eye-tracking; Air traffic controllers; Learning-based methods
• Online Access: https://hdl.handle.net/10356/177834
• Institution: Nanyang Technological University

Air traffic control (ATC) serves a critical role in the aviation industry and is responsible for the safety and efficiency of aircraft movement. High situation awareness (SA), which includes the ability to perceive information (SA level 1), comprehend its meaning (SA level 2), and predict future states (SA level 3), is crucial for air traffic controllers (ATCOs) to manage the airspace safely under varying workload conditions. Eye-tracking (ET) is a prominent physiological method to quantitatively assess operators’ cognitive workload and situation awareness, as eye movements are closely related to one’s attention and information-processing mechanisms. This study aims to investigate the effects of SA levels and workload conditions on ATCOs’ eye movement patterns using ET technology. Experiments were conducted to collect eye movement data of 26 participants while they monitored aircraft on a simulated ATC radar screen and answered freeze-probe queries of different SA levels (level 1 and level 2). Workload conditions were varied by adjusting the number of aircraft participants had to monitor. Two-way repeated measures ANOVA showed significant main effects and interaction effects of SA and workload levels on certain ET metrics (e.g., standard deviation of fixation duration, average fixation duration). These findings could enhance the current understanding of the different SA levels of ATCOs’ tasks and provide more evidence of the relationship between eye movements and cognitive processes. Furthermore, this study integrated learning-based methods with ET data to assess the amount of SA an ATCO possessed. Several conventional machine learning and deep learning models were trained using eye movement data and the models’ classification of controllers’ SA (binary classification of low or high SA) were evaluated. Notably, a hybrid model of a convolutional neural network (CNN) and long short-term memory network (LSTM) (referred to as CNN-LSTM) had optimal performance. The CNN-LSTM model learnt features from the temporal sequential eye movement data and achieved an accuracy and F1 score of 92.7% and 90.2% respectively. The effectiveness and empirical implications of this learning-based method can contribute to future works of developing real-time SA monitoring systems to detect loss of SA in ATCOs, with the aim of reducing human errors and improving aviation safety.