Directed Visibility based Probabilistic Roadmaps

Abstract:
Roadmaps are one of the most widely used techniques within robot path planning. One of the core challenges with generation of roadmaps for planning is that the required number of nodes increases rapidly, with the complexity of the configuration space, quickly yielding very large roadmaps. Techniques such as visibility roadmaps have been introduced for limiting the number of nodes. The past formulations of these do not take into account the possibility of having directional constraints, which require the roadmap to be defined as a directed graph. This work presents a new formulation of the visibility roadmaps that constructs a directed acyclic graph, which are able to facilitate a wider range of planning problems than the traditional visibility roadmaps.