Developing robotic swarms for ocean surface mapping

Abstract

In this paper we describe a model for achieving collective motion in a swarm of robotic boats with the purpose of ocean surface mapping. We also present the implementation and experimental validation of the concept through sea trials with 16-24 robotic boats. The model uses only local neighbor-neighbor interaction and reacts to external information measured in an environmental property or internal information possessed by a leader. We model and demonstrate three kinds of the behaviors: size-tunable mapping, environment-responsive migration, and leader-responsive migration. We also present experimental results from field trials carried out in Catalina Harbor, Santa Catalina Island, California.