To fully utilize the abilities of current autonomous vehicles, it is necessary to understand the interactions between the vehicles and their operators. Since the current state of the art of autonomous vehicles is partial autonomy that requires operators to perform parts of the driving task and to be alert and ready to take over full control of the vehicle, it is necessary to know how the operators' abilities are impacted by the amount of autonomy present in the system. Autonomous systems have known effects on performance, cognitive load, and situation awareness, but little is known about how these effects change in relation to distinct, increasing autonomy levels. It is also necessary to consider these abilities with the addition of secondary tasks due to the appeal of using autonomous systems for multitasking.

The goal of this research is to use a web-based virtual reality study to model operator situation awareness, cognitive load, driving performance, and secondary task performance as a function of five distinct, increasing levels of partial vehicle autonomy first with a constant, low rate of secondary tasks, and then with a steadily increasing rate of secondary tasks. The study had each participant operate a virtual military vehicle in one of five possible autonomy conditions while responding to questions on a communications terminal. After a practice phase for familiarization, each participant took part in two drives where they would have to intervene to prevent crashes regardless of autonomy level.

For both phases, the factors of scored driving performance, secondary task performance, subjective situation awareness, and cognitive load, were analyzed in terms of how they related to the autonomy level and to each other. Results are presented in the form of statistical analysis and modeled equations and show the potential for optimal multitasking within specific autonomy levels and task allocation requirements.

Keywords
AUTONOMY,HUMAN FACTORS,HUMAN PERFORMANCE,MODELING,VIRTUAL

Additional Keywords