Modelling of Robotic Hands: Kinematic and Dynamic Parameter Identification

One of the features that characterizes human versatility is the capability of object complex manipulation. Humans owe part of this capability to the mechanical design and sensory system of our hands. With the advanced performance of human hands on manipulation tasks, researchers have started to develop manipulators that mimic human hands. In order to develop controllers for such manipulators and integrate them into robotic devices it is useful to develop models of the system. This work presents the development of a kinematic and dynamic model of robotic hands. In an experimental setup, a motion capture system is used to obtain finger motions of a pair of prosthetic hands. From the motion capture data and data from the hand sensors, an estimate of the centers and axes of rotation of the finger joints that represent the kinematics of the hands. Furthermore, a dynamic model is derived from data related to the current and the motion of the actuators in the hands. The validity of the models is assessed showing the advantages and future lines of research that this model opens.