Three-Axis Flat and Lightweight Force/Torque Sensor for Enhancing Kinesthetic Sensing Capability of Robotic Hand
Sungwoo Park, Donghyun Hwang
Abstract
When using a robot hand to grip objects, visual information is frequently adopted to determine the external characteristics of objects. However, in reality, purely visual data are insufficient to grasp unspecified objects; somatosensory feedback, which includes tactile and kinesthetic feedbacks, is required to understand the diverse intrinsic properties of objects. We implement a three-axis force/torque (F/T) sensor to measure the interaction force at the joints of a robotic hand and determine the various intrinsic characteristics that aid in grasping previously unknown objects. We primarily focus on a modular, compact, and lightweight structure to embed the sensors in the robotic hand. A novel radially symmetric diaphragm flexure structure is designed to achieve this objective. Several experiments are conducted to validate the performance of the F/T sensor. The functionality of the kinesthetic feedback is demonstrated using a stiffness and weight measurement experiment. The experimental findings demonstrate that the proposed sensor-embedded robotic hand can measure stiffness and weight with errors of 0.0073 N/mm and 8.72 g, respectively.