Investigating what level of visual information inspires trust in a user of a highly automated vehicle

The aim of this research is to investigate whether visual feedback alone can affect a driver's trust in an autonomous vehicle, and in particular, what level of feedback (no feedback vs. moderate feedback vs. high feedback) will evoke the appropriate level of trust. Before conducting the experiment, the Human Machine Interfaces (HMI) were piloted with two sets of six participants (before and after iterations), to ensure the meaning of the displays can be understood by all. A static driving simulator experiment was conducted with a sample of 30 participants (between 18 and 55). Participants completed two pre-study questionnaires to evaluate previous driving experience, and attitude to trust in automation. During the study, participants completed a trust questionnaire after each simulated scenario to assess their trust level in the autonomous vehicle and HMI displays, and on intention to use and acceptance. The participants were shown 10 different driving scenarios that lasted approximately 2 minutes each. Results indicated that the 'high visual feedback' group recorded the highest trust ratings, with this difference significantly higher than for the 'no visual feedback' group (U = .000; p = <0.001 < α) and the 'moderate visual feedback' group (U = .000; p = <0.001 < α). There is an upward inclination of trust in all groups due to familiarity to both the interfaces and driving simulator over time. Participants' trust level was also influenced by the driving scenario, with trust reducing in all displays during safety verses non-safety-critical situations.