How Driving Experience Affect Trust in Automation from Level 3 Automated Vehicles? An Experimental Analysis

Trust in automation is a critical determinant for the acceptance and use of automated systems. Therefore, it constitutes an essential research area to promote the wide application of automated vehicles. A simulator study was conducted with 34 participants to investigate the influence of driving experience (novice, medium experienced and experienced) on trust in Level 3 automated vehicles in different take-over time budgets (7s and 10s) and traffic density (10vehicles/km and 30vehicles/km). Participants encountered four take-over situations and were allowed to freely decide when and how to engage in their mobile phones during simulated driving, in this way to simulate naturally non-driving-related tasks (NDRT) in daily life. Researchers found that driving experience had a significant impact on trust. With the increase of driving experience, participants reported a higher trust level. But in terms of monitoring strategy, experienced drivers still paid more attention to the road or instrument cluster rather than phone, while novice drivers with lower trust were in the out-of-the-loop state more frequently. In addition, higher traffic density and smaller takeover time budget also led to a certain degree of decline in the trust. Results showed that trust in driving automation cannot be fully inferred from the driver's monitoring strategy. The habit of highly focusing on vehicle and road conditions developed by experienced drivers during natural driving would not be easily affected by the trust in automated driving. In addition, novice drivers are more often distracted even though their trust in the system is relatively low, causing greater potential safety hazard. Therefore, in the future design of automated driving, manufacturers should pay more attention to the trust and the out-of-the-loop state of novice drivers, and provide different information interaction for users with different driving experience to ensure appropriate monitoring behavior and trust. In addition, it is necessary to comprehensively analyze the trust in driving automation from multiple levels, combining subjective and objective methods.