Hierarchical Fingertip Space for Synthesizing Adaptable Fingertip Grasps

The ability to synthesize and execute fingertip grasps are bases for dexterous in-hand manipulation. Reliable fingertip grasping is difficult to achieve due to noise and uncertainties in object and hand model, as well as hand control etc. Moreover, in many cases it is desirable to employ an adaptive approach that can deal with changed external forces. In this paper, we propose an approach to jointly optimize stability, adaptability, and reachability of grasps using combinatorial optimization for a hierarchical representation of the fingertip space. To illustrate our approach, we demonstrate an example synthesized by the proposed framework and executed by an Allegro hand. We also show how it is adapted when a perturbation is applied.