Suchergebnisse

Objective: We investigated expertise differences in pilot decision making by examining a hypothesized attention-action link. During simulated flight we measured the accuracy and latency of more and less expert pilots' decision outcomes and used eye tracking to measure their attention. We also examined whether decision outcomes and attentional strategies were influenced by properties of the cues indicating problems. Background: Errors in decision making contribute to many accidents and incidents, especially among inexperienced pilots. Although much is known about decision errors in terms of their outcomes, less is known about the cognitive processes that underlie expert differences in aviation decision making. Method: Fourteen more expert and 14 less expert pilots flew 16 flights in a simulator. Half of the flights contained a failure requiring diagnosis and action in response to the failure. Cues signaling the failures varied in how diagnostic and/or correlated they were. Results: The more expert pilots generally made better decisions in terms of speed and accuracy. Both groups made faster correct decisions when cues were high versus low in diagnosticity. Only the more expert pilots made faster correct decisions when cues were correlated. More attention was allocated to relevant cues (measured by percentage dwell time on areas of interest) when a failure was present, primarily among expert pilots. Moreover, the amount of attention to cues was associated with decision accuracy. Conclusion: The findings support the link between greater attention and more effective decision making. Application: The expert advantage in attention underlying decision outcomes may provide targets for improving pilot training.

In the first part of the reported research, 12 instrument-rated pilots flew a highfidelity simulation, in which air traffic control presentation of auditory (voice) information regarding traffic and flight parameters was compared with advanced display technology presentation of equivalent information regarding traffic (cockpit display of traffic information) and flight parameters (data link display). Redundant combinations were also examined while pilots flew the aircraft simulation, monitored for outside traffic, and read back communications messages. The data suggested a modest cost for visual presentation over auditory presentation, a cost mediated by head-down visual scanning, and no benefit for redundant presentation. The effects in Part 1 were modeled by multiple-resource and preemption models of divided attention. In the second part of the research, visual scanning in all conditions was fit by an expected value model of selective attention derived from a previous experiment. This model accounted for 94% of the variance in the scanning data and 90% of the variance in a second validation experiment. Actual or potential applications of this research include guidance on choosing the appropriate modality for presenting in-cockpit information and understanding task strategies induced by introducing new aviation technology.