Smart Soft Robotic Manipulator for Artificial Intelligence of Things (AIOT) Based Unmanned Shop Applications

Abstract

A smart soft robotic manipulator using a triboelectric tactile (TT) sensor for contact detecting and a triboelectric bending (TB) sensor for deformation monitoring is developed with (Internet of things) IoT module to collect the data and machine learning (ML) algorithm to fuse the output of two types of sensors, realizing grasped objects recognition with high accuracy of 97.321% for 28 different shapes of objects. The time-related features hidden in the triboelectric spectrum, i.e. contact sequence, collision vibration, etc., other than the individual physical information calculated from the output signal, greatly enhance the recognition ability of the integrated intelligent system. By leveraging the IoT and artificial intelligence (AI) analytics, an augmented reality (AR) enabled product preview system is successfully implemented based on the AIoT platform, showing the potential of the integrated system in unmanned shop applications.