Flock of Robots with Self-Cooperation for Prey-Predator Task

Abstract

In this paper, we present a way to lead a swarm of robots through four parameters called repulsion, attraction, orientation, and influence, which are inspired by the behavior of biological societies. Considering the kinematics and dynamics of the robots, we made computational simulations to test the swarm performance and to know the impact of parameters for a prey-predator task. The methodology was experimentally tested in a flock of implemented robots, despite hardware and software limitations. We propose the capture time and statistical metrics to quantify the swarm performance. The results of experimental implementations are consistent with computational simulations based on the robot kinematics and dynamics. Cooperation emerges between the predators while trying to catch the prey, and the change of parameters allows governing on the formation and behavior of the swarm in a decentralized way. Some potential applications for this task include protection, rescue, capture, among others.