부산대학교 도서관

Swarm and modular robotics are an emerging area in control of autonomous systems. However, coordinating a large group of interacting autonomous agents requires careful consideration of the logistical issues involved. In particular, inter-agent communication generally involves time delay, and bandwidth restrictions limit the number of neighbors with which each agent in the swarm can communicate. In this paper, we analyze coherent pattern dynamics of groups of delay-coupled agents, where the communication network is an Erdös-Renyi graph. We show that overall motion patterns for a globally-coupled swarm persist under decreasing network connectivity, and derive the bifurcation structure scaling relations for the emergence of different swarming behaviors as a function of the average network degree. We show excellent agreement between the theoretical scaling results and numerical simulations.