A flexible wearable sensor based on anti-swelling conductive hydrogels for underwater motion posture visualization assisted by deep learning

Abstract

The use of conventional wearable sensors in underwater settings is often impeded by issues such as water swelling, reduced conductivity, and poor adhesion, hindering the progress of underwater sensing technologies. In this study, a double network hydrogel was developed by combining bacterial cellulose (BC) with a copolymer of acrylic acid (AA) and sulfobetaine methacrylate (SBMA). This hydrogel, leveraging the combined effects of hydrophobic association and electrostatic interactions, demonstrated exceptional anti-swelling properties. The presence of numerous hydrogen bonds and dynamic coordination bonds within the hydrogel network conferred remarkable stretchability (>1304%), high toughness (1.3 MJ m⁻³), and high sensitivity (GF = 2.14). Wearable sensors utilizing this hydrogel were able to precisely and consistently capture real-time motion signals from various environments, including air, underwater, and seawater. Employing a two-dimensional convolutional neural network (2D-CNN) deep learning algorithm to integrate and analyze underwater swimming data, the sensors accurately identified and classified 16 swimming postures with a recognition accuracy of 99.37%, offering a novel solution for safety alerts and postural adjustments during underwater activities. This research introduces innovative approaches for developing high-performance wearable sensors for underwater applications, with promising applications in intelligent sensing and human–computer interaction fields.