Game theory based task planning in multi robot systems

In the paper we discuss a problem of planning and coordination in a multi robot system. A team of robots that performs a global task in a human-made workspace of complex structure is considered in this paper. A hybrid architecture of the team motion control system is considered. The system can be split into two layers: the planner module and the behavior based collision free motion controller, that is designed to perform several elementary navigation tasks. The role of the planner is to plan and coordinate execution of elementary tasks by individual agents to obtain performance of global task. The method of elementary tasks planning based on Nperson game is presented. The Nash equilibrium concept of the game solution is applied. An algorithm of multi robot workspace exploration is presented as an example of application of the proposed method. Simulation of the algorithm was carried out, and its result is presented and discussed in the paper.