Uncertainty and heuristics for underactuated hands: grasp pose selection based on the ppotential grasp robustness metric

Abstract The Potential Grasp Robustness (PGR) metric considers different states of the contact points, relaxing requirement all points being far from friction cone boundary. addition new for each point increases computational complexity, which is combinatorial on number and takes a long time grasping configurations with large hand-object contacts. In this work we analyse complexity two recently proposed heuristics, consider that: (i) minimum needed could be in (ii) an analysis grasp data provides most common combinations that lead to accurate estimation PGR. For selecting configurations, PGR computation approach not robust because assumes measured forces at do have uncertainty. propose uncertainty based metric, coefficient variation selection experiments show similar results pose metric. uses find more stable than maximization conventional Article highlights Development heuristics metrics underactuated hands. Definition uncertainty-based se- lection Reduction reality gap physics-based hands