The Effect of Driving Automation on Drivers’ Anticipatory Glances

In this paper, we report a secondary analysis of data collected from two driving simulator experiments to understand the effects of SAE-Level 2 driving automation on drivers’ glances in anticipation of traffic events. Background: Current state-of-the-art consumer vehicle automation requires drivers to monitor the road and intervene when automation fails. Limited research has investigated the effects of automation on drivers’ anticipation of upcoming traffic events. We recently reported two driving simulator studies that focused on drivers’ glance behaviors before such events; however, we did not compare the results of these two studies. Methods: In this paper, we report statistical analyses comparing the glance data from these two studies that had 32 participants each, half of whom were novices and the other half were experienced drivers. The two experiments were comparable in terms of the driving scenarios that required anticipation: the first experiment focused on driving without automation; while the second focused on driving with automation consisting of adaptive cruise control and lane keeping assistance. Further, half of the participants in each experiment were provided with a self-paced visual-manual secondary task. Results: In the no-secondary-task condition, drivers in the automation experiment spent a higher percent of time glancing at anticipatory cues that indicated an upcoming traffic event than did drivers in the no-automation experiment. In the secondary-task condition, no such difference was observed between the two experiments. Conclusion: When there is no distraction to engage in, it appears that automation can allow drivers to have increased visual attention to anticipatory cues.