An efficient deterministic sequence for sampling-based motion planners

This paper presents a deterministic sequence that has the following good and useful features for sampling-based motion planners. It generates samples over a hierarchical grid structure in an incremental low-dispersion manner, allowing a uniform coverage of the configuration space. Due to its grid structure, neighborhood relationship between samples is easily computed, thus allowing roadmap-based planners to faster construct the roadmap. The sequence is computationally efficient and permits to locally control the degree of resolution required at each region of the configuration space by allowing the generation of more samples where they are most needed. The proposed sequence has been incorporated to a PRM-like planner and tested on a bend-corridor problem for different degrees of freedom