Incremental Grid Sampling Strategies in Robotics

Abstract

We present algorithms for generating deterministic sample sequences using incremental grid-based sampling. Our algorithms are designed to generate dense sample sequences over spaces common in robotics, such as the unit cube, SO(3), and SE(3). Our sampling techniques provide the advantageous properties of uniformity, lattice structure, and incremental quality. In addition, the inherent structure of grid-based sequences not only enables them to be used in the place of other sampling techniques in existing algorithms, but also permits the development of new algorithms aimed at exploiting this structure.