On the need for communication in distributed implementations of LTL motion specifications

We revisit the problem of automatic deployment of robotic teams from temporal logic specifications over regions of interests in the environment. In our previous work, we developed an algorithm that could accommodate arbitrary communication constraints, but had two main limitations: (1) it only allowed for communicating robots to move, and (2) it was computationally very expensive. In this paper, we present two approaches to address these limitations. First, we show that if identical robots are allowed to communicate for all times, then the computation is cheaper. Second, we develop an algorithm to test if a given global specification can be implemented by the robots without the move-only-when-communicate constraint.