Design of Hybrid Fully Actuated and Self-Adaptive Mechanism for Anthropomorphic Robotic Finger

Abstract

Prior research on robotic hands predominantly focuses either on high degree of freedom of fully-actuated fingers to replicate a real human hand or on creative designs of under-actuated fingers to make a self-adaptive motion. However, in most cases, fully-actuated fingers encounter difficulty in grasping unstructured objects while under-actuated fingers experience problems in performing precise grasping motions. To deal with any possible scenarios, this study presents a novel design of an anthropomorphic robotic finger that combines both advantages – fully-actuated and self-adaptive grasping (aka FASA) modes – at once. Actuated by tendons, the FASA finger can grasp objects adaptively and achieve accurate angle positioning with the same mechanical design which is introduced in this paper. Based on the kinetostatic analysis, the guideline for the selection of torsion spring is proposed to fulfill the functions of the FASA finger and attain the optimal design of torsional stiffness which manifests itself in a series of tests on different configurations of torsion spring. Likewise, the kinematic analysis for the fully-actuated mode is given as the proof that two joints can move independently by controlling two motors. Ultimately, experimental results again reflect the capability of the FASA finger to perform not only independent precision angle motion but also self-adaptive grasping motion without any change in mechanical structure.

Bibliometric analysis and visualization of research and development trends in humanoid robotics

“…Controlling the motion of robots also a very hot topic, Seyed Adel Alizadeh Kolagar proposed Learning from Demonstrations, a pioneering technique that allows robots to learn to generate 3D motion from 2D data [62], achieving similar results with expensive sensors at a much lower cost, and it can be predicted that this technique will be widely used in motion control with fewer data, for example, Younghyo Park's design of an iPad drawing robot, in which he generates the motion of a robotic arm through video learning [63]. The dexterous movements of manipulators are often driven through high degrees of freedom; Chun-Tse Lee proposed an improved gripper to reduce the need for degrees of freedom, and his results demonstrated the usability of FASA fingers for adaptive motion [64]. Robotic applications also include rescue efforts for disasters, and Xin Shu designed Dexbot, a track-legged humanoid robot that can adapt to a variety of ground environments [65].…”