CONTROL OF MULTI-ROBOT TEAMS BASED ON LTL SPECIFICATIONS

In this paper we address the problem of planning and controlling a group of point robots from specifications given in terms of LTL formulas over regions of interest of a polytopal environment. Our approach consists in three main steps. First, the problem is discretized by abstracting the motion capabilities of each robot. Second, a transition system abstracting the motion capabilities of the whole team is constructed, and a discrete solution to the problem is found by using an approach similar to model checking. Third, the discrete solution is adapted back to the continuous world. The main challenge in this adaptation is the design of a synchronization strategy for the whole team. The approach is fully automated and illustrative examples are included.