AI-Enabled E-Skin with High-Accuracy Material and Texture Recognition via Stick-Slip and Contact Electrification

Bionic electronic skin (e-skin) could endow intelligent robot hands and prosthetics with the ability to perceive materials and texture. Herein, an artificial intelligence (AI)-enabled self-powered e-skin with high-accuracy material and texture recognition via a comprehensive sensing strategy of stick-slip and contact electrification (SC e-skin) is proposed in this work, providing a brand-new solution for human-machine interaction interface. SC e-skin generates stable output voltage based on the coupling of triboelectric and electrostatic induction, and the self-powered sensing was realized subsequently. To make the best of the stick-slip, the macroscopic structure of the interdigital silver electrode and micro structure of hair-like fluorinated ethylene propylene (FEP) surface were fabricated simultaneously, constructing the compact and flexible bionic SC e-skin with polyimide (PI) substrate. The differences of the output signal generated by SC e-skin may not be noticed by human vision clearly, therefore, artificial intelligence was introduced to recognize the material type and even the surface texture. Benefiting from machine learning, the recognition accuracy of six different materials and four kinds of carbon steel with various roughness could achieve as high as 94.88 % and 95.75 %, respectively. This work could pave the way for soft robotics in digital twin and metaverse applications.