Design and Control of a Multi-grasping Bionic Transradial Prosthesis

Abstract

The high number of degrees of freedom (DoFs) of human hand introduces difficulty to design an effective prosthetic hand that enables its user to perform the regular activities of daily living. In that sense, the aim of this research is to investigate the minimum number of actuators required to design and control prosthetic hand. Such problems involve dimensionality reduction of human hand kinematics. Recent studies have used quantitative or qualitative techniques either (a) to reduce the number of hand DoFs only or (b) to suggest actuating strategies for prosthetic hand. However, movements coordination that involving several hand joints have not been investigated yet. In that sense, this study tries to group hand joints by investigating each joint movement coordination during performing different grasp tasks. To accomplish this, multi linear regression (MLR) models are constructed to describe the movement coordination for each hand joint during performing each grasp task. Then, hand joints are clustered into different groups by analyzing the similarities between joints’ movement coordination across grasp tasks using hierarchical cluster analysis (HCA) technique. Finally, the reduced number of DoFs are assigned to each group. The hand dimensionality is reduced by using Principal Component Analysis (PCA). The results suggest that (a) hand joints can be either grouped into six groups that can be derived by thirteen actuators or all joints are grouped into one group and be driven by eight actuators , (b) prosthetic hand may need to utilize fingers abduction/adduction movements, (c) coupling of joints strategy need to be based on joints topology, and (d) insights on hand joint movements’ coordination. Key words: Kinematic Reduction, Degrees of Freedom, Prosthetic Hand.