Using eye tracking to support professional learning in vision-intensive professions: a case of aviation pilots

Abstract In an authentic flight simulator, the instructor is traditionally located behind the learner and is thus unable to observe the pilot’s visual attention (i.e. gaze behaviour). The focus of this article is visual attention in relation to pilots’ professional learning in an Airbus A320 Full Flight Simulator. For this purpose, we measured and analysed pilots’ visual scanning behaviour during flight simulation-based training. Eye-tracking data were collected from the participants (N = 15 pilots in training) to objectively and non-intrusively study their visual attention behaviour. First, we derived and compared the visual scanning patterns. The descriptive statistics revealed the pilots’ visual scanning paths and whether they followed the expected flight protocol. Second, we developed a procedure to automate the analysis. Specifically, a Hidden Markov model (HMM) was used to automatically capture the actual phases of pilots’ visual scanning. The advantage of this technique is that it is not bound to manual assessment based on graphs or descriptive data. In addition, different scanning patterns can be revealed in authentic learning situations where gaze behaviour is not known in advance. Our results illustrate that HMM can provide a complementary approach to descriptive statistics. Implications for future research are discussed, including how artificial intelligence in education could benefit from the HMM approach.