RONA: A Clustering Decision-Making Framework in Robotics Swarm

Local interaction is a salient feature of robotics swarm. How to make decisions under local interaction has been paid large attention. Previous decision-making methods in robotics swarm either contain a centralized center or robots own the ability to communicate with others globally. Once this center is damaged or the communication globally assumption is not satisfied. Swarm can be misled by partial information and make chaotic decisions. This paper proposes a clustering decision-making framework, termed RONA, to address the robotics swarm decision-making problem under local interaction condition. It includes two models i.e. clustering and intra-cluster group decision-making(GDM). The clustering structure divides robots into cluster heads, cluster members, and cluster gateway, which can make the swarm stable and orderly. As for the group decision-making model, it can help swarm robotics make full use of neighbor information under partial communication. As a proof of concept, we implement it into a target entrapping task with an inter-cluster negotiation model among cluster heads. The experiments show the feasibility, extensibility, and superiority of our framework.