Deep‐Learning‐Assisted Thermogalvanic Hydrogel E‐Skin for Self‐Powered Signature Recognition and Biometric Authentication

Abstract

Abstract Self‐powered electronic skins (e‐skins), as on‐skin human‐machine interfaces, play a significant role in cyber security and personal electronics. However, current self‐powered e‐skins are primarily constrained by complex fabricating process, intrinsic stiffness, signal distortion under deformation, and inadequate comprehensive performance, thereby hindering their practical applications. Herein, a novel highly stretchable (534.5%), ionic conductive (4.54 S m −1 ), thermogalvanic (1.82 mV K −1 ) hydrogel (TGH) is facilely fabricated by a one‐pot method. Owing to the formation of Li + (H 2 O) n hydration structure, the TGH presents excellent anti‐freezing and non‐drying performance. It remains flexible and conductive (3.86 S m −1 ) at −20 °C and shows no obvious degradation in the thermoelectrical performance over 10 days. Besides, acting as a self‐powered e‐skin, the TGH combined with deep learning technology for signature recognition and biometric authentication is successfully demonstrated, achieving an accuracy of 92.97%. This work exhibits the TGH‐based e‐skin's tremendous potential in the new generation of human‐computer interaction and information security.