Rearranging multiple objects is a critical skill for robots so as to
effectively deal with clutter in human spaces. The challenge arises
from combinatorially large, continuous C-spaces involving multiple
movable bodies and complex kinematic constraints. This work proposes
hierarchical frameworks for efficiently solving prehensile object
rearrangement with a robotic manipulator. Useful rearrangement
primitives are identified for solving certain classes of problem
instances, which are then composed by higher-level task planners.
Using more powerful primitives, which are reasoning about the
underlying combinatorial and multi-object nature of the problem, are
shown to be beneficial in the context of such hierarchical schemes.