Using an Integrated Cognitive Architecture to Model the Effect of Environmental Complexity on Drivers’ Situation Awareness
The goal of this research is to computationally model and simulate drivers’ situation awareness (SA). In order to achieve this, we have developed a computational cognitive model in a cognitive architecture that can be connected to interact with a driving simulator, as means to infer quantitative predictions of drivers’ SA. We demonstrate the theory of modelling and predicting SA through the lens of human cognition utilizing the QN-ACTR (Queueing Network-Adaptive Control of Thought-Rational) framework as a foundation. We integrate a dynamic visual sampling model (SEEV) to create QN-ACTR-SA in order to allow the model to simulate realistic attention allocation patterns of human drivers. A driver model is also incorporated within QN-ACTR-SA architecture that can simulate human driving behavior by interacting with a driving simulator with the help of virtual modalities such as motor, visual and memory functions. A preliminary validation study is conducted to determine whether SA results of the model correspond to empirical data. The model is probed with SA queries similar to how a Situation Awareness Global Assessment Technique (SAGAT) is conducted on human participants. A comparative assessment demonstrates the model’s ability to simulate drivers’ SA in both easy (with fewer traffic vehicles and signboards) and complex (with more traffic vehicles and signboards) driving conditions.