Estimating Cognitive Load and Cybersickness of Pilots in VR Simulations via Unobtrusive Physiological Sensors

Comparing Presence, Workload, and Performance in Desktop and Virtual Reality Flight Simulations

“…would increase the explanatory power to predict the pilot's workload. Pupil diameter change has previously been associated with Duty Cycle and Aggressiveness derived from stick deflections in VR flight tasks [23]. Recently, EEG data have been implemented in an offline classification of workload of pilots during real flight [36].…”

“…C. Procedure). It is common within the field of aviation to have a small sample size, due to limited access to resources, such as skilled pilots, instructors, and simulator time (e.g., [9], [20], [21], [23]). Besides, our sample is very homogeneous, as the subject-level variation in the dataset is (near) zero (cf., section II.…”

“…Previously, Rajshekar Reddy et al [23] attempted to estimate the workload of non-pilot participants using PIW and physiological measures in a VR flight simulator, when flying a combination of six maneuvers. They found that Duty Cycle and Aggressiveness differed between the tasks, and observed that Aggressiveness correlated with subjective ratings of workload, while Duty Cycle correlated with the number of eye saccades.…”

“…They found that Duty Cycle and Aggressiveness differed between the tasks, and observed that Aggressiveness correlated with subjective ratings of workload, while Duty Cycle correlated with the number of eye saccades. Based on these findings, [23] concluded that a two-dimensional measure of PIW can give an indication of the complexity of the flight maneuvers and hence can be employed in evaluating pilots' workload and performance in a non-invasive manner. However, more research is required to further understand the impact of workload manipulations on two-dimensional (i.e., Duty Cycle and Aggressiveness) as well as one-dimensional forms of PIW, for example, by studying the relationship between mental workload and PIW within-tasks rather than using different maneuvering tasks for PIW measurement.…”

“…Each pilot conducted 12 trials of the speed change task where workload in half of the trials was manipulated by adding a secondary task which was an auditory 2-back task [24]. Using this design, we intended to measure changes in PIW that could be fully attributed to the changes of workload, rather than task characteristics as in [23]. To do this, we extracted Duty Cycle, Aggressiveness, and one-dimensional PIW from the stick data, and compared these measures between low and high workload trials.…”