Bi-modal synergistic hydrogel sensors coupled with machine learning enable gesture parsing and identity recognition

Flexible, wearable epidermal electronic devices from conductive hydrogels have garnered considerable attention because of their seamless skin integration, enabling real-time health monitoring, gesture recognition and haptic feedback, thus facilitating more natural human–machine interactions (HMI). However, most existing conductive hydrogel-based HMI systems often rely on single-modal sensors, limiting the diversity of data features and hindering the extraction of rich and comprehensive information extraction. Moreover, constrained by the stretchability and adhesion of conductive hydrogels, their recognition accuracy requires enhancement in complex and dynamic environments. In this work, we successfully develop a novel conductive hydrogel (PAT-NWs) through a simple immersion method. The resulting hydrogel exhibits strong adhesion (>100 kPa), high stretchability (>2581 %), rapid self-healing and outstanding transparency (86 %). By integrating the exceptional resistance sensing and electrophysiological monitoring capabilities of the PAT-NWs, the constructed bi-modal gesture recognition system achieves 98.33 % accuracy in decoding various gestures. Furthermore, a smart glove integrating dual-modal sensors is designed to enable real-time manipulator control during the execution of different gestures. Building upon machine learning, we subsequently develop an innovative intelligent identity recognition system incorporating a fusion convolutional neural network (CNN), achieving 100 % accuracy with consistent passwords while addressing challenges such as environmental variations and password leakage. This work provides fresh insights for the evolution of life-like tactile systems and HMI interaction, thereby advancing the integrated development of multifunctional flexible perception.