User-driven preferred driving-style guidance interaction for personalizing autonomous driving style

Abstract

Matching the autonomous driving style to its user’s preference is core to a satisfactory user experience. The recent HCI community has undertaken a significant amount of research to understand user-preferred driving styles in autonomous vehicles (AVs). From such studies, it is shown that the true value of autonomous driving-style customization is not fulfilled when the AV learns how the users drive, but is fulfilled when it learns how the users actually want to be driven. Taking into account that users of fully autonomous vehicles still want to be involved even when they do not necessarily have to, a human-machine collaboration system may suggest a feasible means that can support the user-centered input process of their driving preferences in AVs. This thesis aims to design and suggest preferred driving-style guidance as a novel interaction method for personalizing autonomous driving styles. In this interaction, AV passengers guide their driving preferences to AV agent, in a will to adjust autonomous driving style to match their needs. To design and materialize the interaction, this thesis includes a literature review, an on-road Wizard-of-Oz study, an FGI-inspired expert workshop, and an interface design workshop. The first phase of the research composed of theoretical and empirical findings suggests the possible benefits and hindrances of preferred driving-style guidance interaction. Through two types of workshops, we developed principles and guidelines for designing the guidance interaction to achieve a high level of human control while preventing excessive control of humans in the life-critical domain of autonomous driving. The findings, discussions, and scenarios derived from this research suggest insights for the novel form of human-machine collaboration.