Trust me – I’m AutoCAB’: Using natural language interfaces to improve the trust and acceptance of level 4/5 autonomous vehicles
Year: 2018
Type: conference-paper
Abstract: A simulator study explored the use of a natural language interface (NLI) to improve the trust and acceptance of level 4/5 connected/autonomous vehicles.Twenty-three participants undertook three journeys, each lasting approximately 7 minutes.Journeys were framed as different 'Taxi of the future' scenarios, with the vehicle, serving as both the taxi and driver/assistant, transporting participants autonomously to a predefined destination.Participants were provided with one of two different interfaces ('traditional visual' and a NLI), designed to maximise passenger comfort, provide locally informed information and resolve journey-related problems encountered on route.The traditional interface communicated information visually using a touchscreen, whereas the NLI interacted with passengers using 'natural' spoken language.Interactions were limited to one type of interface for each journey, with experiences counterbalanced.For the purpose of the study, the NLI was created using a Wizard-of-Oz technique using an actor to mimic the natural language system.Given free-choice to select their preferred interface during the third drive, twenty participants chose the NLI.Subjective acceptance ratings revealed that participants considered the NLI as significantly more useful, also assigning significantly higher ratings of satisfaction and confidence compared to its traditional, visual counterpart.However, trust ratings were largely equivalent for both interfaces.Overall, results show that the NLI was preferred over traditional visual interfaces in an automated driving environment, supporting its use it future autonomous vehicles.Further work could explore how to adapt the linguistic interactions of natural language systems to also foster increased trust amongst users.
Institution: University of Nottingham
Language: English
Subfield: Human Factors and Ergonomics
Field: Social Sciences
Domain: Social Sciences
Open Access status: green
Funder: Innovate UK
OpenAlex ID: W7222366575