Mental Models in Motion: A Framework for Real-Time Evaluation of Drivers’ Understanding of ADAS Features

Abstract

Advanced driver assistance systems (ADAS) can enhance driving safety, yet their effectiveness depends on drivers’ accurate understanding of system capabilities, limitations, and interactions. This study proposes a structured evaluation framework that combines state diagrams, Cognitive Task Analysis (CTA), and behavioral modeling to assess drivers’ mental models of overlapping ADAS functionalities. Twelve participants operated a vehicle equipped with multiple longitudinal and lateral features while providing concurrent verbal protocols. Behavioral indicators—such as pedal hovers and response delays—were recorded and analyzed to identify patterns of confusion and automation misinterpretation. Preliminary findings highlight specific system transitions and interface inconsistencies that contribute to cognitive mismatches and inappropriate driver responses. By linking observed behaviors to conceptual misunderstandings, this framework supports the systematic identification of usability challenges in ADAS design. The approach provides actionable insights for refining automation logic, improving human-machine interfaces, and informing educational strategies aimed at promoting accurate mental models and safer interactions with automated driving technologies.