Spying with a pilot's eye: Using eye tracking to investigate pilots' attention allocation and workload during helicopter autorotative gliding.

Flight safety in helicopters is a critical aspect of overall aircraft operational safety management, particularly during engine failures requiring autorotative glide, which makes it extremely challenging for the pilot to land the helicopter successfully. In this study, we evaluated the workload and attention allocation of helicopter pilots under such circumstances. In the experiment, a helicopter flight simulator was used to simulate level flight followed by autorotative glide, with the two phases divided into time segments for data collection. First, the data were visualized using heat maps and saccade sequence diagrams, while changes in eye movement metrics (such as peak value and standard deviation) were statistically analyzed. Finally, the criteria through the inter-criteria correlation (CRITIC) method was used to calculate the weight coefficient for each area of interest. This evaluation system was further applied to analyze and compare the changes in eye-movement data and attention to areas of interest during the two phases. The results revealed a shorter fixation duration, but a greater fixation number during the autorotative glide phase. Further, the mean pupil diameter changed over a larger range than during level flight (in level flight, the mean was 5.229 mm, while the standard deviation was 0.059 mm; in autorotative glide the corresponding values were 5.326 mm and 0.126 mm, respectively). For the tachometer, the weight coefficient matched the color of the heat map (2.7 % and colorless during level flight, but 23.8 % and red during autorotative glide), while those for the airspeed indicator and forward view differed significantly between the two phases. This discrepancy stemmed from the fact that during autorotative glide, the pilots prioritized monitoring aircraft rotation speed and attitude, with a particular focus on the forward view, rotor speed, and airspeed, resulting in a more concentrated attention distribution compared to that achieved during level flight. These results confirmed a significant increase in pilot workload during autorotative glide landing, while a shift was observed from low-frequency long gaze time during level flight to high-frequency short gaze time during autorotative glide. Furthermore, the pilots allocated 81 % of their attention to the tachometer, airspeed indicator, and forward views. Adopting this strategy can improve pilots' landing success and provide flight students with valuable training advice to prevent landing failures when helicopters lose power.