This paper presents a study of visual scanning patterns of novice and expert air traffic controllers, executed in a simulator at the Air Traffic Control Centre in Malmö.

Knowledge about how visual information is used in the work of Air Traffic Control (ATC) is important in order to design the user interface (which is now going through a major update in many European countries), to design the air traffic controller education and in order to develop efficient training tools for rated controllers. Eye movements studies could also contribute to the understanding of what characterises ATC work in terms of cognitive aspects.

The purpose of air traffic control is to keep aircraft at a safe distance from one another and to make them fly the shortest possible path, in other words to make the air traffic safe and efficient. To obtain these goals, air traffic controllers use cognitive skills such as planning, problem solving, memory and perception. Combining these skills in a successful manner results in a good so-called situation awareness (SA). Controllers have to distribute their attention to a number of perceptual sources, most of which are visual (radar screen, flight strip board, weather information screen) although some concern other modalities (neighbouring controllers, pilots).

Method: Four novice controllers and four experienced controllers each worked in a simulated 30 min scenario with a medium workload. The interface with radar screen, strip board and communication devices is a typical example of the interfaces now slowly being replaced. A head mounted eye tracker was used to measure the subjects eye movements and pupil diameter.

The results show that the expert controllers spent a significantly larger portion of time looking at the radar screen than the novice controllers did. There is, however, no significant difference between the groups regarding how much the subjects looked at different 'areas of interest' on the radar screen. Instead there are considerable individual differences between controllers within the groups.

To pay more attention to the radar screen is likely to increase subjects' SA. That the expert controllers spent more time looking at the radar screen can probably be explained by the fact that the novice controllers were not experienced in using the flight strips, and therefore had to look more at them. We can further explain this difference in terms of the experts being capable to chunk more flight strip information in each glance than the novices were.

Both experienced and novice controller exhibited very individual search strategies, in terms of how many times controllers look at different areas of interest, i.e. areas on the radar screen that contain important information or areas where the probability of conflicts between aircraft is high. The reason for these differences, and whether they implicate different SA levels, remains to be investigated.