Effects of Robot Control Mode on Situation Awareness and Performance in a Navigation Task

Abstract

Our ability to complete tasks in remote environments has improved with increased sophistication of robotic systems. The quality of remote task performance is driven by the quality of human interaction with robots which can be dependent upon the operator's ability to acquire situation awareness (SA) for task completion. There are, however, often critical limitations in development of SA, which are projected to have a magnified effect when operators must task and control multiple unmanned systems. In this study, we investigated the effects the mode of robot control on operator performance and ability to develop SA while navigating two robots through two mazes. Participants were randomly assigned to groups to control the robots in serial (one after another) or in parallel (at the same time) either manually or using two types of automation. The participants also viewed a systems interface to monitor the status of the robots. Results indicate that participants under serial control produced higher SA scores, though not significantly different from SA scores observed under parallel control modes. The best navigation performance was observed under parallel control with a high level of autonomy that followed a prescribed set of navigation rules. This control mode also resulted in the lowest workload scores. Participants were better able to correctly respond to SA queries for lower level, perceptual items in the robotics task as compared to higher level comprehension and projection items. Performance in the monitoring task was found to be significantly, positively correlated to participant SA.