Creating High-Quality Roadmaps for Path Planning

A central problem in robotics is planning a collision-free path for a robot in a static environment with obstacles. Such a path is usually extracted from a roadmap that represents the connectedness of the free space. Contemporary algorithms generally do not give guarantees on the quality of the roadmap, and, hence, the quality of the paths. While techniques exist for optimizing the paths in the on-line query phase, they are too slow to be applied in real-time applications. In this paper we present a technique for creating high-quality roadmaps, shifting the optimization process to the off-line construction phase. As a result, high-quality paths can be extracted from these roadmaps at interactive speeds. Our method creates roadmaps from which paths can be extracted satisfying the following criteria. First, we require that a path for a query can always be found if one exists (at a given resolution). In addition, we require low query times and low memory consumption. These requirements are met by the Reachability Roadmap Method (rrm), which is an efficient new algorithm that creates small, resolution complete roadmaps for problems involving two or three degrees of freedom such as mobile robots. Then, we require that short and alternative paths can be obtained which is met by adding useful cycles to the roadmap created by the rrm. When safety is of major importance, we require that paths can be obtained having a large clearance to the obstacles. This criterion is met by retracting the local paths corresponding to the edges of the roadmap to the medial axis. Since all operations are performed in a preprocessing phase, high-quality paths can now be extracted from the roadmap in real-time.