Conformal Bioadhesive Hydrogel-Based Multimodal Wristband for Accurate, Mobile, and Real-Time Sign Language Translation and Human-Machine Interactions

Hydrogel-based electrophysiological sensors have garnered copious research interest for wearable human-machine interaction (HMI), whereas existing HMI interfaces suffer from lack of self-adhesion, mismatch of mechanical properties with skin, insufficient accuracy, mobility, and real-time performance. Here, smart hydrogel epidermal electrodes based on honey and sodium chloride co-modified polyacrylamide composite are proposed with superior self-adhesion, skin compliance, stability, conductivity, antibacterial property, and biocompatibility. The hydrogel electrodes are integrated into a flexible adaptive sensing wristband, forming a conformal, stable, and low-impedance interface with biological tissue. The wristband enables the stable acquisition of both dual-channel high-frequency surface electromyography (sEMG) signals and low-frequency acceleration signals with a high signal-to-noise ratio (31 dB for sEMG). By combining a multimodal signal fusion algorithm with a lightweight and efficient Artificial Intelligence model, the wristband achieves a sign language recognition translation accuracy up to 97.2%. Furthermore, the study develops a dedicated mobile application (APP) in which the fusion algorithm and AI model are locally deployed and executed in real time on mobile devices. The integrated standalone HMI system demonstrates excellent accuracy, real-time response, mobility, and extendability in performing various wireless intelligent interaction tasks under complex environments, including sign language recognition, robotic control, active rehabilitation, and motion monitoring, without using the internet.