Drivers' Attention to Preview and Its Momentary Persistence

Participants used a position control system to track the center of a simulated winding roadway with preview that ranged from 0.3 to 1.0 s. Participants' spatial distributions of attention were measured by perturbing the roadway with different frequency sinusoids at different roadway positions and then measuring the degree to which those frequencies were present in their tracking movements. Consistent with Miller's optimal control theory analysis of tracking with preview, participants exhibited a continuous range of attention, and it lengthened with the amount of displayed preview. When the displayed preview disappeared for 5 s, the time to reach a steady-state level of tracking error based only on feedback control was measured. More displayed preview was strongly correlated with longer times to regress to feedback control. One interpretation of this finding is that when preview is withdrawn, visual sensory memory of the previewed roadway can be used for a fraction of a second to prolong the period of feedforward control. Attention can be shifted to relevant positions of the sensory memory image to anticipate the roadway curvature. However, the quality and usefulness of sensory memory gradually diminishes until the participant can only perform feedback control. When preview was restored, the time to reach steady-state feedforward control was correlated with the range of attention. These findings characterize some of the dynamic aspects of attention when drivers are briefly interrupted from viewing the upcoming roadway. This measurement technique may be useful for characterizing spatially distributed attention in other active control contexts.