Designing User Interfaces for Fully Autonomous Vehicles: to Speak, Tap or Press?

Abstract

Fully autonomous vehicles will herald in a new era of personal mobility. However, the transition from manual to self-driving cars could be difficult for many. Improving the user experience will help manage the transition, increase adoption, and realise the benefits of fully autonomous technology. Hence, this thesis investigates how a positive user experience can be delivered and which interaction technologies can contribute to the design of a fully autonomous vehicle interface. The theoretical foundation of the literature review draws from the fields of human-machine interaction, autonomous vehicles, and human-robot interaction. The Design Research Methodology was used as a framework by which to plan and execute this research. First, an exploratory sequential mixed-methods study investigated the users’ needs and expectations of autonomous vehicle interfaces. It consisted of both a qualitative interview study and a quantitative online survey which found that users wanted different information to be shared with them, and a different interior, depending on if the vehicle was privately owned or shared. This resulted in a dichotomy of interfaces for each ownership model. Second, two simulator experiments investigated three interaction technologies for user interfaces. Findings indicated that subjective metrics, especially satisfaction and usefulness, were better predictors of improvement, when compared to purely relying on performance metrics. While a touchscreen-only interface was more performant, an interface that included a voice assistant was found to be both more satisfying and more useful. The results of this research have led to the development of a list of design recommendations about how the user experience of fully autonomous vehicles can be improved and which interaction technologies may be employed in providing improvements. Ultimately, this work concludes that while “conventional” automotive interface technologies are mostly sufficient to control a fully autonomous vehicle, a voice assistant and other novel technologies can provide additional improvements in user experience.