A Random Glance at the Flight Deck: Pilots' Scanning Strategies and the Real-Time Assessment of Mental Workload

Abstract

Based on previous research showing the usefulness of spatial statistics in detecting randomness in the distribution of eye fixations, this study investigated the ocular behavior of professional pilots engaged in a simulated flight. The distribution of eye fixations was used as an indirect index of mental workload: Eye movements were recorded during the different phases (departure to landing) of a simulated flight and were analyzed using spatial statistics algorithms. Results showed sensitivity of spatial dispersion indices to variations in mental workload: higher during departure and landing, lower during climb and descend, and the lowest during the cruise phase. This finding provides additional evidence of the utility of fixations distribution as a real-time measure of mental workload and, consequently, as a trigger for adaptive automation.

Fixation identification: The optimum threshold for a dispersion algorithm

“…Although this index was previously used in eyetracking to investigate the dispersion of fixations in a recording (Camilli et al, 2008;Di Nocera, Camilli, & Terenzi, 2007;Di Nocera, Terenzi, & Camilli, 2006; , the present study is interested in the dispersion of PORs within fixations. Since the tremor, drift, and microsaccades are in- for different participants (West, Haake, Rozanski, & Karn, 2006).…”

Gaze entropy reflects surgical task load

“…In non-surgical scenarios, gaze entropy-a measure of the uncertainty over the gaze position at any point in time [32] (see ''Gaze dynamics analyses: Gaze entropy [and velocity]'' section for more details)-has been proved to be a sensitive and valid task load index [33][34][35]. Briefly, gaze entropy increases with task load [34]: that is, visual exploration patterns become less stereotyped (i.e., more random) during more complex tasks.…”

“…Briefly, gaze entropy increases with task load [34]: that is, visual exploration patterns become less stereotyped (i.e., more random) during more complex tasks. Furthermore, in medical applications, gaze entropy has been successfully used to recognize the rising and falling interest levels of expert radiologists during medical imaging assessment and, then, track their learning curves [25].…”

“…When compared with other eye metrics, such as pupil response [21,27,36], gaze entropy offers the advantage that the measured value does not depend on external factors such as environmental light or emotional state [37,38]. Furthermore, unlike other sensitive indices, as saccadic (fastest movement of the eye) velocity [19], gaze entropy does not rely on intrusive and bulky eye tracking systems for the acquisition of its signals [34,39]. Thus, gaze entropy could be measured with mobile and wearable eye trackers (e.g., small-framed glasses eye trackers) which are robust to head movements and changes in ambient illumination, and, most importantly, are easy and comfortable to use.…”

Understanding Optimal Decision-Making in Wargaming

“…For example, opposing cognitive states, i.e., engaged vs. relaxed, normal vs. distracted, and fatigued vs. alert, were modeled from eye movement and pupil size (Marshall, 2007); other research demonstrates that eye-tracking measurements can be used to determine level of visual processing load (Van Orden et al, 2001). Along with others, our work with helicopter pilots has shown that visual scan pattern can be used to detect underlying cognitive strategies that the pilots may not even be aware of using (Di Nocera et al, 2007;Yang et al, 2011). Below we describe relevant findings regarding eye tracking and domain specific expertise, and how the combination of eye tracking technology combined with reinforcement learning theory can enable us to understand how visual scan patterns change as expertise is developed.…”