Modelling the Relationship Between Pilot Visual Scan, Control Activity, and Aircraft State

The visual scan of pilots is theorized by first principles to have a close relationship with aircraft state and pilot control activity. This paper presents modelling techniques that capture and represent the visual scan of pilots as categorical time series models that account for the pilot’s recent visual gaze and control activity, as well as the aircraft’s state. By appropriately conditioning data, creating Task-Relative datasets, and then Vector Auto-Regression modelling, insight into the pilot’s visual scan gain be gained. Highlighting the extent that these models identify attributes of visual scan that have only been hypothesized hitherto, the models are shown to be similar when conducting similar tasks but change with task evolution.