SYDE Graduate Seminar - Cognitive Architecture Modelling for Understanding Pilot Behaviour: Applying QN-ACTR to Flight Training in Takeoff and Landing
Presenter
Robin Xu, PhD candidate in Systems Design Engineering
Abstract
Flight training assessment should capture not only maneuver completion but also how control actions unfold over time. We extend prior takeoff-focused work to model both takeoff and landing using QN-ACTR in closed-loop interaction with X-Plane. The model uses stage-specific targets, threshold-based transitions, and bounded corrective actions to support interpretable process-level assessment. We evaluated it against data from 10 pilots flying a Cessna 172N at Brantford Airport. Comparisons used maneuver duration, rotation and touchdown speeds, vertical-speed variability, RMSE, and MAPE. The model reproduced complete maneuver traces and broadly comparable milestone speeds, but showed different pacing and lower variability, especially during landing. The results demonstrate the feasibility of cognitive modelling for process-level flight assessment and highlight the need to better capture perturbations and individual differences.
This seminar counts towards the graduate student seminar attendance milestone!