In this paper we introduce a task-based method for designing underactuated multi-joint prosthetic hands for specific grasping tasks. The designed robotic hands or prosthetic hands contain fewer independent actuators than joints. We chose a few specific grasping tasks that are frequently repeated in everyday life and analysed joint motions of the hand during the completion of each task and the level of participation of each joint. The information was used for the synthesis of dedicated underactuated mechanisms that can operate in a low dimensional task coordinate space. We propose two methods for reducing the actuators� number. The kinematic parameters of the synthesized mechanism are determined by using a numerical approach. In this study the joint angles of the synthesized hand are considered as linearly dependent on the displacements of the actuators. We introduced a special error index that allowed us to compare the original trajectory and the trajectory performed by the synthesized mechanism, and to select the kinematic parameters of the new kinematic structure as a way to reduce the error. The approach allows the design of simple gripper mechanisms with good accuracy for the preliminary defined tasks.
Loading....