The instructor assistant system (iASSYST) - utilizing eye tracking for commercial aviation training purposes.

This work investigates the potential of providing commercial aviation flight instructors with an eye tracking enhanced observation system to support the training process. During training, instructors must deal with many parallel tasks, such as operating the flight simulator, acting as air traffic controllers, observing the pilots and taking notes. This can cause instructors to miss relevant information that is crucial for debriefing the pilots. To support instructors, the instructor ASsistant SYSTem (iASSYST) was developed. It includes video, audio, simulator and eye tracking recordings. iASSYST was evaluated in a study involving 7 instructors. The results show that with iASSYST, instructors were able to support their observations of errors, find new errors, determine that some previously identified errors were not errors, and to reclassify the types of errors that they had originally identified. Instructors agreed that eye tracking can help identifying causes of pilot error. Practitioner summary: This paper introduces an instructor assistant system, which is evaluated in a user study involving 7 airline flight instructors. The system can be used by airline flight instructors to complement their observations, as a basis for discussions with pilots during debriefing, and by airline pilots to improve their flight performance.

The effect of flight phase on electrodermal activity and gaze behavior: A simulator study.

Current advances in airplane cockpit design and layout are often driven by a need to improve the pilot's awareness of the aircraft's state. This involves an improvement in the flow of information from aircraft to pilot. However, providing the aircraft with information on the pilot's state remains an open challenge. This work takes a first step towards determining the pilot's state based on biosensor data. We conducted a simulator study to record participants' electrodermal activity and gaze behavior, indicating pilot state changes during three distinct flight phases in an instrument failure scenario. The results show a significant difference in these psychophysiological measures between a phase of regular flight, the incident phase, and a phase with an additional troubleshooting task after the failure. The differences in the observed measures suggest great potential for a pilot-aware cockpit that can provide assistance based on the sensed pilot state.