Abstract: As the Internet of Things advances, gesture recognition emerges as a prominent domain in human-machine interaction (HMI). However, interactive wearables based on conductive hydrogels for individuals with single-arm functionality or disabilities remain underexplored. Here, we devised a wearable one-handed keyboard with gesture recognition, employing machine learning algorithms and hydrogel-based mechanical sensors to boost productivity. PCG (PAM/CMC/rGO) hydrogels are composed of polyacrylamide (PAM), sodium carboxymethyl cellulose (CMC), and reduced graphene oxide (rGO), which function as a strain, pressure sensor, and electrode material. The PAM chains offer the gel's elasticity by covalent cross-linking, while the biocompatible CMC improves the dispersion of rGO and promotes electromechanical properties. Integrating rGO sheets into the polymer matrix facilitates cross-linking and generates supplementary conductive pathways, thereby augmenting the gel system's elasticity, sensitivity, and durability. Our hydrogel sensors include high sensitivity (gage factor (GF) = 8.18, 395.6 %-551.96 %) and superior pressure sensing capabilities (Sensitivity (S) = 0.3116 kPa^-1, 0-9.82 kPa). Furthermore, we developed a wearable keyboard with up to 98.13 % accuracy using convolutional neural networks and a custom data acquisition system. This study establishes the groundwork for creating multifunctional gel sensors for intelligent machines, wearable devices, and brain-computer interfaces.

Key words: Gesture recognition, Hydrogel, Mechanical sensors, Human-machine interaction