Autonomy beyond Control: Theory-based Control Allocation for Vehicle Interface Design

Abstract

Research has proposed a variety of cooperative systems that allow drivers to participate in the driving task. While previous work shows that such interfaces can improve user experience, trust, and acceptance, reasons for their effectiveness remain unclear. This paper investigates the effect of different components of autonomy. By operationalizing positive liberty (condition to act upon self-endorsed values) and negative liberty (condition to act without external restrictions) in a vignette study (N=155), we show that differences between automated and manual driving do not emerge from the inherent properties of these concepts but mostly from perceived positive liberty, which seems to be the strongest predictor of willingness to use and psychological need fulfillment. Task allocation and study results indicate that drivers value being included in tactical and strategic decision-making over operational control. These findings suggest that designing for decision authority, rather than control, is key to user acceptance in automated driving.