Sampling-based Control Synthesis for Multi-Robot Systems under Global Temporal Specifications

This paper proposes a sampling-based algorithm for multi-robot control synthesis under global Linear Temporal Logic (LTL) formulas. Robot mobility is captured by transition systems whose states represent regions in the environment that satisfy atomic propositions. Existing planning approaches under global temporal goals rely on graph search techniques applied to a synchronous product automaton constructed among the robots. As the number of robots increases, the state-space of the product automaton grows exponentially and, as a result, graph search techniques become intractable. In this paper, we propose a new sampling-based algorithm that builds incrementally a directed tree that approximates the state-space and transitions of the synchronous product automaton. By approximating the product automaton by a tree rather than representing it explicitly, we require much fewer resources to store it and motion plans can be found by tracing the sequence of parent nodes from the leaves back to the root without the need for sophisticated graph search techniques. This significantly increases scalability of our algorithm compared to existing model-checking methods. We also show that our algorithm is probabilistically complete and asymptotically optimal and present numerical experiments that show that it can be used to model-check product automata with billions of states, which was not possible using an of-the-shelf model checker.