Neural Networks for Multi-Finger Robot Hand Control

Abstract. this paper investigates the employment of artificial neuralnetworks (ann) for a multi-finger robot hand manipulation in whichthe object motion is defined in task-space with respect to six cartesianbased coordinates. the approach followed here is to let an ann learnthe nonlinear functional relating the entire hand joints positions anddisplacements to object displacement. this is done by considering theinverse hand jacobian, in addition to the interaction between handfingers and the object being grasped and manipulated. the developednetwork has been trained for several object training patterns andpostures within a cartesian based palm dimension. the paperdemonstrates the proposed algorithm for a four fingered robot hand,where inverse hand jacobian plays an important role in robot handdynamic control