Simplified Human Hand Models for Manipulation Tasks

Abstract

The human hand is the most dexterous and versatile biomechanical device that possesses the human body, this device created by Nature during millions years of evolution represents one of the more distinctive qualities among other animals. Since the 70s and 80s, important contributions have appeared in physiological studies of the human hand (I. A. Kapandji, 1970), (I. A. Kapandji, 1981). Studies about robotic hands have made several contributions e.g. the Stanford/JPL hand (K. S. Salisbury & B. Roth, 1983), the Utah hand (S. C. Jacobsen et al., 1986), the Okada hand (T. Okada, 1982), the Belgrade/USC hand (G. Bekey et al., 1990), the UB hands (C. Melchiorri and G. Vassura 1992), (C. Melchiorri and G. Vassura 1993), the DLR hands (J. Butterfass et al., 1999), (J. Butterfass et al., 2001), the University of Tokyo hand (Y. K. Lee & I. Shimoyama, 1999), Barrett Hand (W. T. Townsend, 2000), the Robo-Naut hand by NASA (C. S. Lovchik et al., 2000), the Karlsruhe University ultra-light hand (S. Schulz et al., 2001), the GIFU hand (H. Kawasaki et al., 2001), the Shadow Dextrous Hand (Shadow Robot Company), (F. Rothling et al., 2007), a prosthetic hand (H. Yokoi et al., 2004), the DLR-HIT-Hand (H. Liu et al., 2008) and other. These devices have different kinematic configurations with respect to the number of Degrees of Freedom (DoF) controlled, number of fingers, number of joints, type of actuation, etc. This chapter describes simplified human hand models that properly represent the kinematic behaviour of the human hand in accordance with the precision and application required. The first part describes a human hand model with 24 DoF. This model represents a balance between complexity and realism. Simplified human hand (SHH) models are analyzed using the model with 24 DoF. These SHH models (1 to 24 DoF) are evaluated in accordance with the level of dexterous or power required. A Cyberglove® (Immersion) is used for the experiments carried out in this work. Kinematic constrainswere checked with the information provided by the glove. Also, this glove was used for evaluating the error of the SHH versus the full 24 DoF hand model. Finally, the experiments carried out with SHH and 24 DoF hand model compare the efficiency in grasping for circular and prismatic grasps in accordance with the application. 10