Biomimetic Neural Intelligent E-Skin System for Tactile Perception and Robotic Decision-Making

Abstract

The widespread application of electronic skin (e-skin) in human–machine interaction necessitates intelligent and information-rich systems. However, the rapid and efficient deployment of e-skin for high-precision multisensor fusion remains a critical challenge. This study introduces a pioneering biomimetic neural intelligent e-skin system that significantly enhances human–machine interaction and robotic perception capabilities. Our innovative approach integrates two novel e-skin technologies: a highly flexible multiwalled carbon nanotube (MWCNT) based e-skin for precise pressure sensing, and a gallium–indium alloy liquid metal e-skin with exceptional stretchability for motion capture. The MWCNT e-skin, fabricated through a simple carbon nanotube impregnation method, achieves ultrathinness (<1 mm), ease of preparation, and inherent flexibility. The liquid metal e-skin, developed using a unique dispersion and reconstruction method, exhibits excellent linearity (R2 > 99.9%) and impressive stretchability (∼700%). By integrating our two types of e-skins, our system has achieved multidegree-of-freedom control and tactile feedback for robotic arms. It demonstrates the capability to perform object grasping tasks solely through tactile feedback in visually challenging environments, including underwater conditions. The system achieves a 98.26% accuracy in identifying diverse objects and making autonomous decisions through tactile sensing alone, showcasing its self-decision-making abilities. This research establishes a new paradigm for intelligent robotics, advancing human–machine interaction in complex environments.