Abstract
We present a model of where humans place index finger and thumb when grasping arbitrary objects with a precision grip. The model incorporates three weighted rules to (a) maximize force-closure, (b) optimize the position of the object's gravicenter relative to the fingers, and (c) minimize deviations from ones natural grasp angle. We determined the parameters of the model in two experiments (N=18 and N=19) with objects of simple geometry. In a third experiment (N=18) we predicted the grasp points for a set of objects with complex geometries. The results show that our simple model can surprisingly well predict human grasp point selection.
This work was supported by the DFG Graduiertenkolleg GRK 885 “NeuroAct”.