" Human Factors Evaluation Framework for Telemetry-Grounded Conversational Onboarding in Autonomous Vehicles"

"Autonomous vehicles require structured onboardingmechanisms to support calibrated trust and reduce uncertaintyduring first-time passenger exposure. However, there is currentlyno standardized human factors evaluation framework for as-sessing conversational onboarding systems that are groundedin real-time vehicle telemetry and compared against appro-priate baselines. This paper proposes a structured evaluationframework that combines a telemetry-grounded conversationalarchitecture, a standardized onboarding interaction protocol, acontrolled within-subject experimental design, and an ordinaltrust\u2013anxiety assessment methodology. A reference implemen-tation was developed using a Unity-based virtual autonomousdriving environment, a Python Socket.IO telemetry bridge, anda real-time conversational surface hosted in LiveKit. In an initialvalidation study with nine first-time participants, the frameworkcompared three conditions: no assistant, unidirectional voiceupdates, and bidirectional conversational onboarding. Resultsshowed statistically significant increases in trust and reductions inanxiety following conversational onboarding, with large within-subject effect sizes. Participants also consistently preferred theconversational onboarding condition over unidirectional voiceupdates for comfort, trust induction, and satisfaction. Thesefindings provide initial empirical validation that the proposedframework can sensitively detect onboarding-related trust andanxiety calibration effects and can be applied to systematic eval-uation of conversational interfaces in intelligent transportationsystems."