Sampling-Based Optimal Control Synthesis for Multi-Robot Systems under Global Temporal Tasks

摘要: This paper proposes a new optimal control synthesis algorithm for multi-robot systems under global temporal logic tasks. Existing planning approaches goals rely on graph search techniques applied to product automaton constructed among the robots. In this paper, we propose sampling-based that builds incrementally trees approximate state-space and transitions of synchronous automaton. By approximating by tree rather than representing it explicitly, require much fewer memory resources store motion plans can be found tracing sequences parent nodes without need sophisticated methods. significantly increases scalability our compared existing We also show proposed is probabilistically complete asymptotically optimal. Finally, present numerical experiments showing approach synthesize from automata with billions states, which not possible using standard algorithms or off-the-shelf model checkers.