A Framework for Understanding Trust in Automated Vehicles

As automated vehicles become an important area of focus for automation development, understanding the interactive dyad between driver and vehicle is increasingly important to improve driving efficiency, technology use and uptake, and transportation safety. The level of trust that the driver or other vehicle occupants have towards the vehicle has been shown to be one of the primary aspects that influences the acceptance and adoption of automated vehicles. A lack of trust can lead drivers to reject the use of the system and its potential benefits, while overtrusting the system's capability may increase the risk of road traffic collisions. To calibrate trust appropriately, it is crucial to understand the underlying process of trust development and evolution, followed by an appropriate evaluation of the level of trust. Against this background, this study reports on the findings of a comprehensive review of the literature around driver-vehicle trust in automated and autonomous driving. This review aims to identify the definitions, influencing factors, evolution, and evaluation of trust. This review presents the findings using a conceptual framework to understand how driver-vehicle trust may be established by (a) providing an overview of different definitions of trust in interpersonal, human-automation, human-artificial intelligence (AI), and human-robot contexts as a foundation for redefining trust specifically for driver-vehicle trust in automated and autonomous vehicles; (b) analysing the influencing factors of trust, and dividing them into three top-level aspects: driverrelated, vehicle-related, and situation-related; (c) fundamentally describing the evolution process of driver-vehicle trust with a two-layer model based on static and dynamic aspects; and (d) discussing different subjective and objective methods for evaluating trust in automated driving, and considering further research for an objective evaluation of trust.