An exploration of the applicability of information processing theories in road hazard perception context using e-scooter simulator in augmented virtuality scenarios

This paper aims to provide a cognitive perspective on the hazard perception process in the traffic situation based on the conceptual frameworks of two complementary theories of brain information processing, namely, Signal Detection Theory and Predictive Coding Theory. The study uses an augmented virtuality scenario encompassing a simplified traffic situation, where participants are faced with hazard cues characterised by a different degree of predictability. Throughout the experiment, the acceleration and braking behaviour of an e-scooter rider together with the electroencephalography (EEG) data is collected. The developed experimental setup allows for testing the applicability of brain theories in explaining behaviour and cognitive processing of the perception of potential hazards with the ultimate goal to improve road safety. Current findings support the a priori expectations showing that participants create predictions concerning future potential hazards. Produced predictions and the subsequent behaviour are modulated by the degree of ambiguity of hazard cues in line with Signal Detection Theory. Moreover, following Predictive Coding Theory, the predictions improve as more external input is gathered, and the mental model is updated. Complementary to behavioural results the alpha wave is used as a neural marker of hazard predictability. The results provide implications for road safety researchers and practitioners, where the inclusion of a cognitive perspective can guide the more-informed design of road infrastructure as well as in-vehicle human support systems to be more aligned with the processing mechanisms of human cognition and exploit the synergies between them.