Page 27 - Human Inspired Dexterity in Robotic Manipulation

Digital Hand: Interface Between the Robot Hand and Human Hand 23

Fig. 2.6 Synthesis of power grasp and computation of convex hull of the friction cones.
(A) Synthesized posture. (B) Computed convex hull.

Fig. 2.7 Synthesis of precision grasp and computation of contact forces. (A) Synthesized
posture and computed contact forces from palmar side. (B) Synthesized posture and
computed contact forces from dorsal side.

Fig. 2.6 also shows the convex hull of friction cones computed from the
distribution of contact points with the assumption that the sum of normal
contact forces is equal to one. From this convex hull, the grasp quality that
represents the stability of grasp is derived by computing the radius of the
maximum inscribed sphere [20]. This is quite important because the stability
of grasp proposed in robotics has a positive correlation with the subjective
response, the ease of grasp that subjects feel when they are grasping cylinder
or quadric prisms. If we construct a regression model that relates the stability
of grasp computed from the digital-hand model and subjective response
revealed by questionnaires, no further user experiments are necessary. We
can estimate the subjective response from a digital-hand simulation and
the regression model.