A three-dimensional skin-conformal hybrid electrode system for burden-free sEMG monitoring during exercises

Abstract

High-fidelity surface electromyography (sEMG) acquisition under dynamic conditions is critical for rehabilitation, sports monitoring, and human-machine interaction. Conventional electrodes often suffer from poor compliance, weak adhesion, and unstable skin-electrode interfaces during motion, degrading signal quality. In this paper, we propose a three-dimensional hybrid electrode system (THES) that synergistically integrates a thermo-responsive sponge-gel electrode with a stretchable electronic architecture and wireless sEMG acquisition. The electrode achieves rapid body-temperature-triggered adhesion, high stretchability, and low impedance, enabling conformal contact and stress absorption. Simulations and experiments confirm that the three-dimensional (3D) structure mitigates stress concentration and maintains conductivity under deformation, thereby stabilizing the skin-electrode interface during dynamic motion. The system records sEMG signals at a high sampling rate, supporting muscle force estimation, fatigue monitoring, and motion recognition. In badminton scenarios, THES enables stable on-skin sEMG acquisition during dynamic movements, supporting the recording of distinct muscle activation patterns across different stroke postures as well as long-duration exercise. Leveraging these high-quality signals, our dual vision transformer (DualViT) framework, which fuses dual-modal features, classifies eight stroke actions with an accuracy of 99.34%, outperforming conventional models. By synergistically optimizing materials, structures, and algorithms, THES advances sEMG recording under dynamic conditions, offering a viable platform for wearable bioelectronic applications.

Keywords: Burden-free sEMG monitoring; Dual-modal deep learning; Skin-conformal 3D hybrid electrode; Stretchable flexible circuit; Thermo-responsive hydrogel.