Path planning in dynamic environment using an adaptive mesh

Abstract

Solutions for the common problem of path planning in an abstract environment have been extensively studied in many scientific disciplines. However, many explored techniques assume the environment does not change and that there is a complete and detailed overview of this examined space. In addition, many path planning methods need to derive a specific graph structure from the environment representation and it can be often very difficult to obtain this structure in some real applications.In our paper, we introduce a general model for the real-time path planning in a dynamic environment and provide a hybrid technique that combines a graph and grid representation of the examined space. The proposed path planning method uses an adaptive mesh for its graph part to provide the capability of the assimilation to the changing environment.The presented method offers faster times for the path retrieval than the classical raster based approaches and works in a dynamic environment where the conventional graph based techniques fail. On the other hand, there are still some higher memory requirements of the proposed solution due to the necessary raster representation of the examined environment.