Eye movements as an indicator of situation awareness in a flight simulator experiment

Table of Contents

A possibility to measure situation awareness by eye movements

Methods

Generalization potential

A possibility to measure situation awareness by eye movements

Van de Merwe, Van Dijk and Zon (2012¹) explored the possibility to employ eye tracking to measure situation awareness (SA) of pilots when they have to deal with a malfunction in a simulator cockpit setting.
The result showed that the accuracy rates were higher when there were more eye fixations and distributed more systematically. The successful crew was able to find the solution because of the structured scanning pattern.

	Fixation rate	Dwell time	Entropy
Relationship between normal level, malfunction level and display typet	More Significant	More Significant	Significant
Relationship between search period and ECAM display	More Significant (-)	More Significant (-)	-
Relationship between Nav and ECAM display	More Significant (-)	More Significant (-)	-
A trend between PFD and ECAM	Not Significant	Significant (-)	-

CARS Scores in the Postperiod

Situation Awareness Level	p value
Perception	Significant
Comprehension	Significant
Projection	Significant
Resolution	Significant

Methods

Research approach

This was an exploratory study.

Sample

12 airline pilots, competent to fly Airbus A320, paired up as a captain and a first officer as a crew.

Design and Materials

• The true experimental design was used.
• A flight simulator experimentation was arranged for learning the relationship between SA and eye movements. The simulation facility was configured as an Airbus A320.
• Eye trackers, cameras, and optical head tracker were employed to trail the pilot’s visual scanning behavior in the cockpit.
• Moreover, the eight question posttrial questionnaire with a 4-point scale, CARS was employed to compare the pilot’s interpretation of their SA and the result from the eye tracking.

Procedure

• This study was done in a flight simulator setting. The differences of pilots’ scanning entropy and attention focus were watched while they were searching for the malfunction.
• The hypothetic trip was about 25 minutes from London Heathrow to Amsterdam Schipol Airport.
• The observed areas were PFD, navigation display (NAV), electronic centralized aircraft monitoring display (ECAM) and other inside and outside views in the cockpit.
• Four levels of SA were compared by using CARs outcomes.

Variables

Independent variables:
• System state in normal level
• System state in malfunction level
Dependent variables
• Eye track measure
• Dwell time and
• Fixation rates on the displays
• Enthropy
• A malfunction scenario

Data analysis

• ANOVA was performed on the accuracy of collected data.
• The choice of significance level was measured by alpha 5%.
• A proper effect size for the contrast between two averages and a biased estimator of the difference reasoned by the model in the population were employed.

Generalization potential

• Although the findings are supported by previous research in the relationship between eye movements and SA, the relatively small number of participants restricts the external validity for generalization.
• This research is potentially generalizable to pilots when they have to deal with a malfunction in a simulator setting.
• The study is also applicable to real life situations for pilots, because of the flight simulator was set up as an Airbus A320 cockpit.

References

1. Koen VAN DE MERWE, VAN DIJK Henk & ZON Rolf (2012). Eye movements as an indicator of situation awareness in a flight simulator experiment. International Journal of Aviation Psychology. Jan-Mar2012, Vol. 22 Issue 1, p78-95. 18p.