Synthesizing Cooperative Controllers from Global Tasks of Multi-robot Systems

Abstract

The reactive system synthesis for GR(1) fragment of LTL has been widely studied and used in different works. Meanwhile, automatic procedures for generating distributed systems from global behaviors are also thoroughly studied. In this paper, we present a method that automatically constructs cooperative controllers for multi-robot systems from global tasks specified by GR(1). We combine reactive systems synthesis and distributed systems synthesis to solve global task planning problems for multi-robot systems. In short, we specify global tasks for a multi-robot system which consists of several robots, then our algorithm generates individual controllers for robots and also synthesizes a communication strategy. The communication strategy makes robots as a cooperative system through synchronizing them on environment inputs and finally achieves the goal of completing the global tasks. This paper can be used as a cooperative framework for multi-robot systems.