A Comparison of Policy Search in Joint Space and Cartesian Space for Refinement of Skills
This addresses a specific bottleneck in robotics for skill refinement, but it is incremental as it builds on existing policy search methods.
The paper tackles the problem of refining robot skills transferred from humans via imitation learning, which suffers from reachability issues in Cartesian space due to conventional inverse kinematic solvers, and shows that a configurable approximate solver can accelerate refinement considerably.
Imitation learning is a way to teach robots skills that are demonstrated by humans. Transfering skills between these different kinematic structures seems to be straightforward in Cartesian space. Because of the correspondence problem, however, the result will most likely not be identical. This is why refinement is required, for example, by policy search. Policy search in Cartesian space is prone to reachability problems when using conventional inverse kinematic solvers. We propose a configurable approximate inverse kinematic solver and show that it can accelerate the refinement process considerably. We also compare empirically refinement in Cartesian space and refinement in joint space.