Physiological Correlates of Passenger-Facing HMI Communication in Automated Driving

Passenger-facing vehicle interfaces must inform without disrupting non-driving activities. This study related HMI communication during fully automated driving to passenger physiology and interaction quality during concurrent reading. Twenty-six participants completed a no-communication Baseline drive and one HMI drive with abstract auditory-tactile cues (Simple) or synthesised speech (Extended); 24 contributed physiological data. Electrodermal activity (EDA) was the primary measure, with User Experience Questionnaire ratings. Because HMI variant was assigned between participants, participant-level analyses provided primary condition inference. Mean EDA was lower in both HMI conditions than Baseline, but no contrast was statistically supported (Simple–Baseline −73.1, 95% CI [−160.3, 14.0], p = 0.092; Extended–Baseline −44.7, p = 0.296; Extended–Simple 28.4, p = 0.614). No UEQ scale survived multiplicity correction. Segment-level mixed-effects models (803 segments) reproduced this ordering with narrower intervals reflecting repeated observations, not independent assignments. Participant-level contrasts matched the unit of assignment; segment-level models characterised temporal physiological variation.