Modelling level 1 situation awareness in driving: A cognitive architecture approach

Abstract

The goal of this research is to computationally model and simulate the situation awareness (SA) of drivers. A computational model in a cognitive architecture was developed that can interact with a driving simulator to infer quantitative predictions of drivers’ SA. The model uses the Queueing Network Adaptive Control of Thought-Rational (QN-ACTR) framework as a foundation and integrates a dynamic visual sampling model (SEEV) to create QN-ACTR-SA, which simulates attention allocation patterns of human drivers at SA Level 1 (i.e., perception of critical elements). QN-ACTR-SA also incorporates a driver model that can interact with a driving simulator. A validation study was conducted to determine whether Level 1 SA results produced with the QN-ACTR-SA model correspond to empirical data collected from human drivers (14 participants) for the same tasks. Both QN-ACTR-SA and human participants were probed for SA using two approaches: within-task queries using the Situation Awareness Global Assessment Technique (SAGAT) and post-experiment questions. A comparative assessment demonstrated that QN-ACTR-SA could reasonably simulate drivers’ Level 1 SA for two driving conditions: easy (with few vehicles and signboards) and complex (with dense traffic and signboards). QN-ACTR-SA fit for human SAGAT scores (possible range 0–100) resulted in a mean absolute percentage error (MAPE) of 5.0% and the root means square error (RMSE) of 3.5. Model fit for post-experiment human SA results was MAPE of 6.7% and RMSE of 6.1. Limitations of QN-ACTR-SA as a predictive model and areas of future research are discussed.