Building Trust in Highly Automated and Autonomous Vehicles

Abstract

Trust in highly Automated and Autonomous Vehicles (AAVs) is a topic that has been gaining traction in recent years, across academia, the technology industry, and of course, the automotive industry. A number of automotive OEMs and tech companies across the globe are developing AAVs, with the short-term focus being on SAE Level 2 and Level 3 vehicles, and the longer-term focus being on SAE Level 4 and Level 5 vehicles. Alongside the development of the core technology needed for such AAVs, these companies have been grappling with a key question that will influence the uptake and ultimate success of AAVs; How to get people to trust these vehicles? This paper outlines a series of small-scale projects undertaken by Jaguar Land Rover Research between 2016 and 2018 as part of the UKAutodrive project that attempted to address this question. Across this series of user trials, participants experienced a number of autonomous drives in a low-speed (6mph/10kph) prototype SAE Level 4 autonomous ‘pod’ for up to fifteen minutes at a time, and then asked to rate their trust at the end of each individual drive. Overall, the data shows that Trust is something that can indeed be reliably measured, something that changes/fluctuates over time, and can be undone if the occupant experiences a negative event (e.g. a near miss), with the impact of this event depending on when the occupant experienced it (first trip vs. fifth trip). Finally, we show that it seems to be the number of trips in, or exposures to, an autonomous vehicle rather than the length of time per trip that influences trust, with more, shorter trips (8x4 minute trips) recording higher reported trust compared to fewer, longer trips (4x8 minutes trips).

Keywords: Trust, Trust in autonomy, Autonomous Vehicles, ADAS, Self, driving Vehicles

DOI: 10.54941/ahfe1003807