Sampling behavior while detecting conflicts between linear moving stimuli

Sampling behavior while detecting conflicts between linear moving stimuli

Looijestijn, Anouk (TU Delft Mechanical, Maritime and Materials Engineering)

de Winter, J.C.F. (mentor)
Eisma, Y.B. (mentor)
Dodou, D. (graduation committee)

Delft University of Technology

Mechanical Engineering

2018-10-17

Introduction. Air traffic controllers (ATCo’s) are responsible for a safe and efficient air traffic flow, and therefore, they are required to be excellent in conflict detection. Various studies have uncovered relationships between conflict geometry (e.g., conflict angle) and operators’ abilities to detect conflicts. However, little is known about the underlying perceptive and cognitive processes during conflict detection tasks. Knowledge of these processes could give insight into how ATCo’s could be supported. In order to discover how people look at typical air traffic control (ATC) situations, a simplified ATC scenario was presented to novice participants in which two dots (representing aircraft) moved towards each other.
Methods. The eye movements of 35 participants were recorded during an experiment in which they had to indicate whether a conflict was present or not. Each participant watched 36 different videos with a duration of 20 seconds and different air traffic geometries. The independent variables were: (1) conflict angle (2) configuration (3) closest distance of approach, and (4) discrete vs continuous moving stimuli (2 Hz or 30 Hz).
Results. The results show that continuous moving trials obtained a significantly higher performance score, and more and shorter fixations were found compared to discrete moving trials. Furthermore, in accordance with Neal and Kwantes (2009) and Remington et al. (2000), participants performed significantly better with a conflict angle of 30 degrees compared to larger angles. However, it was found that the performance score with a conflict angle of 100 degrees was lower and the self-experienced difficulty was higher, compared to 150 degrees. The eye movement variables showed a monotonic relation with the conflict angle: when the number of fixations increased, the fixation duration decreased with the higher conflict angles. Furthermore, participants sampled more often from one dot to the other and exhibited less pursuit movement with increasing conflict angles. Moreover, it was found that conflict detection was significantly easier when one of the dots moved diagonally. Finally, the results show that for trials in which no conflict occurred, participants exhibited more fixations and sampled more from one dot to the other, compared to trials with a conflict.
Conclusions. We conclude that novice participants are better at detecting conflicts with continuous moving stimuli compared to discretely moving stimuli. If further research shows the same increase in performance with continuous motion in real ATC, flight radars could be adjusted accordingly. Also, it is concluded that the conflict angle has influence on eye movements. Indications are found that conflict angles close to 0 and 180 degrees are easier for detecting conflicts. Further research with various conflict angles is recommended. Furthermore, indications are found that diagonal movements might be easier for conflict detection. Moreover, we conclude that in our experiment pursuit movements are preferred with vertical movements compared to horizontal movement. Finally, we conclude that participants sample from one dot to the other when the dots are further away from each other, but when they come closer to each other, pursuit movement is often used to follow both dots at the same time.

Collision detection
Eye movements
Eye tracking
Visual attention

http://resolver.tudelft.nl/uuid:e93b1978-e8e5-441d-817a-1a0eceda0a4d

Student theses

master thesis

© 2018 Anouk Looijestijn