Ultra–Thin Joint Torque Sensor With Enhanced Sensitivity for Robotic Application

Abstract

As advanced robotic technologies such as human-robot interaction and automatic assembly processes have emerged, torque sensors have become an essential component for robots. However, commercial torque sensors are not suitable for robotic applications because of their large sizes, heavy weights, narrow options, and high prices. In this letter, we develop a novel capacitive joint torque sensor with ultra-thin structure, high performance, but low cost. To achieve these goals, novel designs have been applied to both the sensing and deformable parts, which are the most important elements of the torque sensor. To obtain high sensitivity, a novel electrode structure called wedge electrode was applied to the sensing part, and a new deformable structure was designed to be ultra-thin and easy to manufacture. Then, the electrode and deformable structures were implemented in a single torque sensor. The developed torque sensor was calibrated based on an artificial neural network (ANN) model and verified to perform high accuracy and sensitivity, and low crosstalk by comparing it with a commercial torque sensor. Finally, a high-performance torque sensor was implemented in ultra-thin size with a diameter of 108 mm and thickness of 13 mm.