P. A. Desmond, P. A. Hancock and J. L. Monette University of Minnesota
The effect of automation on the driver's performance and subjective state is a critical issue in the light of future transportation developments such as Intelligent Vehicle-Highway ,Systems (IVHS). In such systems, in-vehicle navigation and collision avoidance systems are integrated with the control software of the vehicle to automate many aspects of the driving task. Automated systems may have negative consequences for the driver's performance and subjective state, Studies in contexts other than driving have demonstrated some of the adverse effects of automation on performance. For example, Endsley and Kiris ( 1995) showed that individuals who are forced to manually operate a system following a breakdown in automation show impaired performance when compared with individuals who perform the task in a manual mode of operation. Two explanations have been proposed to account for automation-induced performance impairments. Endsley and Kiris ( 1995) argue that such impairments can be explained in terms of a loss of situation awareness (SA). Situation awareness is "the perception of elements in the environment within a volume of time and space, the comprehension of their meaning, and the projection of their status in the near future" ( Endsley, 1988, p.97). Endsley and Kiris ( 1995) propose that individuals who have lower SA are slower to detect difficulties and are slower to focus on relevant features of the system which will allow them to identify the problem and regain manual performance. Operator complacency has also been proposed as a major factor that relates to degraded vigilance in monitoring performance under automation. Studies of aircraft cockpits suggest that performance may be impaired by complacency resulting from the pilot's confidence in automated systems ( Singh, Molloy & Parasuraman, 1993). As well as impairing performance, we may expect automated tasks to elicit stressful subjective reactions in individuals. As Matthews et al. (this volume) point out, automated tasks are similar to vigilance tasks since they usually demand sustained monitoring but infrequent response from the individual. Vigilance tasks have been found to lead to symptoms such as fatigue, boredom and a reduction in motivation (e.g. Scerbo, 1998). This chapter outlines two studies that have focused on the impact of an automated system on drivers' performance and subjective state. Implications of the research for future transportation developments are also discussed.
In the first study, 40 drivers performed both a manual and an automated drive on the Minnesota Wrap-around environment simulator (WES). Both drives lasted for 40 minutes. In the manual drive, drivers had fall control over the vehicle throughout the drive. In the automated drive, the vehicle's velocity and trajectory was under system control. Three perturbing events occurred at early, intermediate and late phases in both drives. In the manual drive, the perturbation took the form of a "wind gust" that caused the vehicle to drift towards the edge of the right-hand driving lane. In the automated drive, the perturbing event was a failure in automation that caused the vehicle to veer towards the edge of the right-hand lane at the same magnitude as the wind gusts. Following a failure in automation, drivers were required to manually control the vehicle for three minutes until the automated driving system became operational again. While the automated system was operational, drivers were instructed to keep their hands off the steering wheel. The driver's lateral control of the vehicle, as indexed by heading error, was the principal performance measure that was recorded following the perturbing events in both drives. Heading error is the mean deviation between the direction of the road and the direction of the