Perception-Based Temporal Logic Planning in Uncertain Semantic Maps

This paper addresses a multi-robot planning problem in partially unknown semantic environments. The environment is assumed to have known geometric structure (e.g., walls) and to be occupied by static labeled landmarks with uncertain positions and classes. This modeling approach gives rise to an uncertain semantic map generated by semantic SLAM algorithms. Our goal is to design control policies for robots equipped with noisy perception systems so that they can accomplish collaborative tasks captured by global temporal logic specifications. To account for environmental and perceptual uncertainty, we extend a fragment of Linear Temporal Logic (LTL), called co-safe LTL, by including perception-based atomic predicates allowing us to incorporate uncertainty-wise and probabilistic satisfaction requirements directly into the task specification. The perception-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 are updated online to adapt to a continuously learned semantic map. We provide extensive experiments to demonstrate the efficiency of the proposed planning architecture.