Automation and Crew Performance: The Importance of Who and What

In the past, “automation” in the cockpit has been associated with both an increase and a reduction in workload, performance, and crew communications. A reason for these apparently contradictory results could be that different types of automation were studied, for example, automation that required little vs. considerable interaction with a machine. The current study investigated whether automation of different task domains for different crewmembers has differential effects. Sixteen pilots formed two-pilot crews. The crews each completed four experimental flights in a PC-based simulator, encountering different levels of automation. Systematically, an automated system for the pilot (a simple autopilot) and one for the copilot (a navigation computer) were made available or were unavailable, alone or in combination. The results showed that the automation manipulations improved task performance only when they were combined. Additionally, the use of the navigation computer reduced communications within the cockpit, while the autopilot was associated with larger contributions by the pilot. Further, mental workload and frustration were lower when the navigation computer was available. Also, perceived effort was lower with the autopilot on. The results of the current study suggest a more detailed picture of automation effects than previously known. Additionally, the results suggest that different types of automation can interact to create unexpected effects that may not be uncovered if one studies automated systems individually.