Sensor-Based Temporal Logic Planning in Uncertain Semantic Maps

This paper addresses a multi-robot mission planning problem in uncertain semantic environments. The environment is modeled by static labeled landmarks with uncertain positions and classes giving rise to an uncertain semantic map generated by semantic SLAM algorithms. Our goal is to design control policies for sensing robots so that they can accomplish complex collaborative high level tasks captured by global temporal logic specifications. To account for environmental and sensing uncertainty, we extend Linear Temporal Logic (LTL) by including sensor-based predicates allowing us to incorporate uncertainty and probabilistic satisfaction requirements directly into the task specification. The sensor-based LTL planning problem gives rise to an optimal control problem, solved by a novel sampling-based algorithm, that generates open-loop control policies that can be updated online to adapt to the map that is continuously learned by existing semantic SLAM methods. We provide extensive experiments that corroborate the theoretical analysis and show that the proposed algorithm can address large-scale planning tasks in the presence of uncertainty.