Supervisory control of multiple robots in dynamic tasking environments

Abstract

A military targeting environment was simulated to examine the effects of an intelligent route-planning agent RoboLeader, which could support dynamic robot re-tasking based on battlefield developments, on the performance of robotics operators. We manipulated the level of assistance (LOAs) provided by RoboLeader as well as the presence of a visualisation tool that provided feedback to the participants on their primary task (target encapsulation) performance. Results showed that the participants’ primary task benefited from RoboLeader on all LOAs conditions compared to manual performance; however, visualisation had little effect. Frequent video gamers demonstrated significantly better situation awareness of the mission environment than did infrequent gamers. Those participants with higher spatial ability performed better on a secondary target detection task than did those with lower spatial ability. Finally, participants’ workload assessments were significantly lower when they were assisted by RoboLeader than when they performed the target entrapment task manually.

Practitioner Summary: This study demonstrated the utility of an intelligent agent for enhancing robotics operators’ supervisory control performance as well as reducing their workload during a complex urban scenario involving moving targets. The results furthered the understanding of the interplay among level-of-autonomy, multitasking performance and individual differences in military tasking environments.

Keywords:

• human–robot interaction
• supervisory control
• level of autonomy
• intelligent agent
• military
• individual differences

Acknowledgements

This research was funded by the U.S. Army Research Laboratory Director's Research Initiative Program and the Safe Operations for Unmanned Reconnaissance in Complex Environments (SOURCE) Army Technology Objective (ATO). The authors would like to acknowledge the contributions of the following individuals: Zhihua Qu, Mark Snyder, Daniel Barber, David Adams, Stephanie Quinn, and William Plew.