November 16, 2001

Research on Airplane Cockpit Automation Leads to Air Safety

The cockpits of commercial airplanes are becoming more and more automated with advances in computer technology.  Planes can almost land themselves if necessary, but the complexity of such automated systems can create problems because of the dynamics of pilot–computer interactions.

Dr. David B. Kaber, assistant professor of industrial engineering at NC State University, and graduate students Melanie C. Wright and Heather L. Warren are studying pilot performance under various levels of cockpit automation.  They hope their studies will lead to increased safety through better pilot awareness of changing aircraft conditions.

Numerous possible modes of automation exist in these sophisticated planes, and each level causes the flight controls to behave differently.  Pilot awareness of the current mode is essential for appropriate decisions as a pilot takes-off, navigates and lands a plane. The pilot must remain aware of changes in altitude, heading and speed that are dictated by automated controls. This awareness of the state of the aircraft and surrounding environment is known as “situation awareness” and has been found to be critical to effective pilot performance.  Pilot situation awareness may vary under different modes of cockpit automation.

Kaber's group is observing pilot–computer interactions through an experiment funded by the National Aeronautics and Space Administration (NASA).  The two research objectives are to identify the modes of cockpit automation that are optimal for pilot situation awareness and ability to control the aircraft and to determine what workload level allows pilots to optimize their responses for safe and efficient aircraft operation.  Identifying these variables will enable Kaber's team to make recommendations for future cockpit design.

Sixteen central North Carolina pilots with experience on various types of aircraft participated in the study.  They used a PC-based flight simulator, modeled after a McDonnell Douglas (MD) 11 cockpit, to pilot a flight and respond to an in-flight approach reprogramming challenge.  The results of the study showed that, in general, when the cockpit was in manual mode pilot performance improved after the challenge, whereas in automated mode pilot performance appeared to worsen, despite the fact that automation decreased the pilot's workload.

According to Kaber, it may be that the increased workload of the manual mode keeps the pilot more involved, aware and in control of the situation.  “A critical question is whether automation is taking the pilot out of the loop,” he said.  “It's absolutely essential to keep the pilot involved, so a key question for this research is whether high-level automation in advance of the approach revision causes performance and situation awareness problems.”

This preliminary research has provided some clues that should help cockpit designers create optimal conditions for aircraft efficiency.  Kaber's future studies will involve Delta Airlines pilots in Atlanta, where his team will conduct a field study using an enhanced version of their simulator and commercial pilots with MD 11 experience.  Georgia Institute of Technology researchers will collaborate on this aspect of the study.

In an increasingly automated world, studies such as Kaber's can help designers understand the complexity of human–machine interactions.  The benefit in this case is increased airplane safety.

-- rudd --

Media Contact:  Linda E. Rudd, 919/515-3848, linda_rudd@ncsu.edu
Technical Contact:  Dr. David B. Kaber, 919/515-3086, dbkaber@eos.ncsu.edu,