Hierarchical path planning of mobile robots in complex indoor environments

TLDR: The FHD* algorithm guarantees the optimality of the global path, and requires considerably less time for the path replanning operations, and can be easily extended to the problem of path planning between different floors or buildings.

Abstract: Inspired by the Hierarchical D* (HD*) algorithm of D. Cagigas (Cagigas, 2005), in this paper we introduce a novel hierarchical path planning algorithm called Focussed Hierarchical D* (FHD*). Unlike the original HD* algorithm, the FHD* algorithm guarantees the optimality of the global path, and requires considerably less time for the path replanning operations. This is achieved by several modifications: (i) optimal placement of the so-called bridge nodes needed for hierarchy creation, (ii) focusing the search around the optimal path, which reduces the search area without loss of optimality, and (iii) introduction of partial starts and partial goals, which further reduce computational time of replanning operations. The FHD* algorithm was tested in a multiroom indoor environment and compared to the original HD* algorithm, non-hierarchical D* algorithm, and Focussed D* algorithm under the same conditions. The FHD* algorithm significantly outperforms other algorithms with respect to the computational time. Furthermore, it can be easily extended to the problem of path planning between different floors or buildings.