Investigating Shifts in Driver Attention and Trust Across Manual and Autonomous Driving: Insights from Eye-Tracking Metrics

As autonomous vehicles (AVs) advance, understanding driver visual attention and trust in various driving modes is crucial for ensuring safety and effective design. This study examined the relationship between trust and visual attention in a simulated Level 3 driving environment, utilizing eye-tracking metrics: fixation duration (FD), fixation count (FC), and time to first fixation (TFF). Thirty-four participants navigated scenarios involving automation, handover, manual, and takeover modes while interacting with a human-machine interface. Results from generalized linear mixed models and multivariate regression revealed that higher trust significantly influenced attention allocation, particularly in automation and handover modes. Trust was associated with increased FD and FC on in-vehicle displays and decreased TFF, especially for tablet-related areas of interest. These findings highlight the nonlinearity of trust in modulating visual engagement, suggesting that overly high trust may lead to reduced attention to monitoring-critical interfaces.