Gaze Point Projection by Marker-Based Localization of Eye-Tracking Glasses for Examining Trust in Automated Vehicles

Abstract:
Human trust in automated systems is a major factor for the acceptance of automated vehicles. In this paper, an approach for marker-based visual localization of eye-tracking glasses in a vehicle is proposed. The estimated pose of the eye-tracking glasses within a 3D model of the vehicle is utilized to infer the 3D gaze point with respect to the vehicle frame by ray casting. The approach is utilized in a Wizard-of-Oz study on a test track to analyze the gaze behavior of drivers in terms of their trust in automated vehicles. Objective gaze-related measures are compared to subjective self-assessments of the participants. Results show, depending on the participant, more or less consistent objective and subjective assessments.