Enabling Human Support of Robot Swarms

Abstract

Autonomous robot swarms may offer efficient and parallel task execution compared to single robot systems. They are also more robust than a single robot, but not free of failures. It is difficult for a human to monitor the swarm, because the state space grows exponentially with the number of robots presenting a very high cognitive load. Moreover, errors in robot swarms propagate non-intuitively and it becomes harder to monitor and debug the swarm. Previous works have looked into enabling control of robot swarms by a human user. We instead present a framework that allows a non- expert user to support the robot swarm in the presence of errors by use of a GUI. It enables monitoring, verification and viewing analysis on both robot and swarm behavior. We believe that this framework is a step forward to moving swarms from lab settings to real-world scenarios.