Examining the Visual Attention of Pairs of Operators during a Low to High Workload Change

Abstract

Dynamic and data-rich domains, like those found in the military, primarily rely on multiple operators’ visual attention. Of interest is to understand how shared visual attention impacts performance when workload changes and whether this informs the adaptation process between collaborators. Ten pairs of participants completed a simulated Unmanned Aerial Vehicle control task under two different workload conditions - first under low workload and then under high workload. The best and worst performing pairs were identified and further analyzed by assessing pairs’ percent gaze overlap and strategy when workload changed. The findings showed the best performing pairs not only had higher levels of percent gaze overlap on average, but also increased their percent gaze overlap as workload increased. Additionally, the best performing pairs engaged in the adaptation process with not only their actions, but their overall visual attention allocation strategy. These findings suggest systems and technology in these domains should allow operators to have access to their collaborator’s visual data, in order to provide the opportunity to adapt and dynamically collaborate under different workload conditions.

How Coordinated Visual Attention on a Target Area of Interest is Impacted by a Change in Workload Over Time

“…Following this guideline, no participants were excluded from the study and the mean data loss was 9.23%. Given, previous work showed that the best performing pairs had higher percent gaze overlap on the target AOI (Devlin et al, 2019), percent gaze overlap was then calculated for each minute of the scenario for the target AOI only. The best performing pairs' percent gaze overlap ranged for each scenario (low workload range: 17.4-33.0%; high workload range: 40.0-63.7%) and were consistently higher than worst performing pairs, (low workload range: 1.0-7.3%; high workload range: 9.9-22.2%).…”

“…This research builds on previous work that examined whether shared visual attention affected the collaboration strategies of successful and unsuccessful pairs managing a workload change (Devlin, Flynn, & Riggs, 2019). Our previous work found the most successful pairs of participants adapted both their task completion and shared visual attention strategies more readily upon experiencing a change in workload.…”

“…The secondary tasks included a rerouting task (avoiding no-fly zones), fuel leak task (maintaining UAV health), and chat message task (responding to chat messages). More details can be found in (Devlin et al, 2019).…”

“…The best performing pairs were the three highest scoring pairs and the worst performing pairs were the three lowest scoring pairs. More details can be found in (Devlin et al, 2019). Dependent variables included the two shared visual attention metrics of the target AOI (i.e., video feed panel; Figure 1b): percent gaze overlap and the mean of the maximum absolute value of the phi coefficient for each minute of both scenarios.…”

Does It MultiMatch? What Scanpath Comparison Tells us About Task Performance in Teams

“…Table 2 shows the point value associated with each task (Devlin et al, 2019). Points were assigned to emphasize the priority of the primary task (i.e., target detection) as well as to convey the severity of incorrectly or not attending to a task (e.g., UAV flies through no-fly-zone ).…”