Iterative Transportation by Multiple Mobile Robots Considering Unknown Obstacles

Abstract

The focus in this paper is on a planning method for an iterative transportation task performed by mobile robots in environments including unknown obstacles. This task requires the acquisition of environmental information, the generation of the appropriate path network based on the acquired information, and the formation of a group of robots on the planned path network. To achieve an efficient method of transportation, a motion planning architecture is proposed that includes three phases, i.e., environmental exploration, path generation, and learning of formation. In the first phase, robots cooperatively explore the environment using a learned visibility graph while transporting. Next, a network of transportation paths consisting of 1- and 2-lane paths is generated using two kinds of configuration spaces. In the final phase, every robot learns a behavior strategy by reinforcement learning to acquire an efficient formation of transportation. The simulation results indicate the effectiveness of the proposed architecture.