Combined top-down and bottom-up design of cooperative multi-robot systems

We propose a formal design framework for cooperative multi-robot systems through combining a top-down mission planning with a bottom-up motion planning. In this work, it is assumed that a global mission is given as regular languages, and that basic motion controllers for robots and an environment description are given. Then, our method can automatically synthesize coordination strategies and control policies for the robot team to accomplish the given mission. A mission planning layer is sitting on the top of our framework, where an assume-guarantee reasoning and learning based approach is applied to decompose the global mission into local tasks according to the capabilities of each robot. With respect to these local tasks, the motion planning for each robot is then solved by composing basic motion primitives that are verified safe by differential dynamic logic (dL). The motion primitive composition is through a Satisfiability Modulo Theories (SMT) solver that searches feasible solutions in face of constraints due to local task requirements and the environment description. Our method can handle changing environments as the motion primitives are reactive in nature which makes the motion planning adaptive to local environmental changes. Furthermore, on-line mission replanning can be triggered by lower motion planning layers once no feasible solutions can be found through the SMT solver. The design framework is illustrated through an automated warehouse example.