Perception‐to‐Cognition Tactile Sensing Based on Artificial‐Intelligence‐Motivated Human Full‐Skin Bionic Electronic Skin

TLDR

Traditional e‑skin lacks human‑brain‑like cognition, limiting its progress toward intelligent applications. The authors propose an AI‑motivated full‑skin bionic e‑skin that mimics human vellus hair, epidermis‑dermis‑hypodermis structures. The system employs triboelectric vellus‑hair bionics for contactless tactile perception and a supercapacitive iontronic epidermis‑dermis‑hypodermis structure coupled with a five‑layer MLP to achieve general intelligent tactile cognition. The FSB e‑skin attains ultrahigh sensitivity (8053.1 kPa⁻¹) and sub‑5.6 ms response, and, with a six‑layer MLP, delivers real‑time material and location cognition that surpasses human capability.

Abstract

Traditional electronic skin (e-skin), due to the lack of human-brain-like thinking and judging capability, is powerless to accelerate the pace to the intelligent era. Herein, artificial intelligence (AI)-motivated full-skin bionic (FSB) e-skin consisting of the structures of human vellus hair, epidermis-dermis-hypodermis, is proposed. Benefiting from the double interlocked layered microcone structure and supercapacitive iontronic effect, the FSB e-skin exhibits ultrahigh sensitivity of 8053.1 kPa-1 (<1 kPa), linear sensitivity of 3103.5 kPa-1 (1-34 kPa), and fast response/recovery time of <5.6 ms. In addition, it can realize the evolution from tactile perception to advanced intelligent tactile cognition after being equipped with a "brain". First, static/dynamic contactless tactile perception is achieved based on the triboelectric effect of the vellus hair bionics. Second, the supercapacitive iontronic effect based structural bionics of the epidermis-dermis-hypodermis and a five-layer multilayer perception (MLP) enable the general intelligent tactile cognition of gesture cognition and robot interaction. Most importantly, by making full use of the FSB e-skin with a six-layer MLP neural network, an advanced intelligent material cognition system is developed for real-time cognition of the object material species and locations via one contact, which surpasses the capability of humans.