Multi-robot path-planning based on implicit cooperation in a robotic swarm

摘要

Material handling in manufacturing environments is often proposed as an application of multi-agent robotic systems. In this paper we argue that reactive agents can perform the conveying task very efficiently as long as they are able to coordinate their displacements and to cooperate in such as way as to optimize their paths. A distributed implementation-oriented architecture is proposed which is based on beacon-identified loading points. The implicit cooperation/ coordination scheme relies on beacon impersonation by the agents, leading to complex group behavior in the robotic swarm. Thanks to this simple scheme, agents are able to perform multi-robot distributed path-planning. We present both the theoretical basis of our architecture and simulation results showing that implicit coordination leads to emergent group behaviros such as stable agent formations and implicit obstacle avoidance. Thus it ensures high performance levels for the conveying system.