ABSTRACT
A series of four experiments was conducted in a simulated military mounted crewstation environment to examine the workload and performance of the robotics operator in either a single-tasking or a multi-tasking environment. In the multi-tasking environment, the operator also had to concurrently perform gunnery and communication tasks. In both tasking environments, the robotics tasks involved managing a semi-autonomous ground robot or teleoperating a ground robot to conduct reconnaissance tasks. Results showed that, in the single-tasking environment, the operator’s performance was significantly worse when the robot had to be teleoperated than when it was semi-autonomous. In the multi-tasking environment, in contrary, the operator’s robotics task was significantly worse when the robot was semi-autonomous, implying the operator’s increasing reliance on automation when under heavy workload of multiple tasks. The operator’s robotics task performance improved when the concurrent gunnery task was automated. However, when the automation was not reliable (i.e., miss-prone or false alarm-prone), the operator’s robotics task performance degraded, and the severity of the degradation was affected by both the type of automation unreliability and the operator’s attentional control. More specifically, for high attentional control operators, false alarm-prone alerts were more detrimental than miss-prone alerts. For low attentional control operators, conversely, miss-prone automation was 294more harmful than false alarm-prone automation. Finally, our data indicated that the operators’ preference of cueing display was related to their spatial ability. Low spatial ability operators preferred visual cueing over tactile cueing, and high spatial ability operators favored tactile cueing over visual cueing.
