States of fatigue are implicated in driver impairment and motor vehicle accidents. This article reports two studies investigating two possible mechanisms for performance impairment: (1) loss of attentional resources; and (2) active regulation of matching effort to task demands. The first hypothesis predicts that fatigue effects will be accentuated by high task demands, but the second hypothesis predicts that fatigue effects will be strongest in "underload" conditions. In two studies, drivers performed a stimulated driving task, in which task demands were manipulated by varying road curvature. In a "fatigue induction" condition, the early part of the drive was occupied by performance of a demanding secondary task concurrently with driving, after which the concurrent task ceased. Post-induction driving performance was compared with a control condition in which drivers were not exposed to the induction. In both studies, the fatigue induction elicited various subjective fatigue and stress symptoms, and also raised reported workload. Fatigue effects on vehicle control and signal detection were assessed during and after the fatigue induction. The fatigue induction increased heading error, reduced steering activity, and, in the second study, reduced perceptual sensitivity on a secondary detection task. These effects were confined to driving on straight rather than on curved road sections, consistent with the effort regulation hypothesis. The second study showed that fatigue effects were moderated by a motivational manipulation. Results are interpreted within a control model, such that task-induced fatigue may reduce awareness of performance impairment, rather than reluctance or inability to mobilize compensatory effort following detection of impairment.

No abstract

A driving simulator study investigated the effect of automation of the driving task on performance under fatiguing driving conditions. In the study, drivers performed both a manual drive, in which they had full control over the driving task, and an automated drive, in which the vehicle was controlled by an automated driving system. During both drives, three perturbing events occurred at early, intermediate, and late phases in the drives: in the automated drive, a failure in automation caused the vehicle to drift toward the edge of the road; in the manual drive, wind gusts resulted in the vehicle drifting in the same direction and magnitude as the “drifts” in the automated drive. Following automation failure, drivers were forced to control the vehicle manually until the system became operational again. Drivers’ lateral control of the vehicle was assessed during three phases of manual control in both drives. The results indicate that performance recovery was better when drivers had full manual control of the vehicle throughout the drive, rather than when drivers were forced to drive manually following automation failure. Drivers also experienced increased tiredness, and physical and perceptual fatigue symptoms following both drives. The findings have important implications for the design of intelligent transportation systems. Systems that reduce the driver’s perceptions of task demands of driving are likely to undermobilize effort in fatigued drivers. Thus, the results strongly support the contention that human-centered transportation strategies, in which the driver is involved in the driving task, are superior to total automation.

No abstract

No abstract