Human management of a robotic swarm

Abstract

This paper proposes a management algorithm that allows a human operator to organize a robotic swarm via a robot leader. When the operator requests a robot to become a leader, nearby robots suspend their activities. The operator can then request a count of the robots, and assign them into subgroups, one for each task. Once the operator releases the leader, the robots perform the tasks they were assigned to. We report a series of experiments conducted with up to 30 e-puck mobile robots. On average, the counting and allocation algorithm correctly assigns 95% of the robots in the swarm. The time to count the number of robots increases, on average, linearly with the number of robots, provided they are arranged in random formation.