A multifunctional hydrogel enabled by cellulose-Zn2+ solvation structure for convolutional neural network assisted motion recognition.

Integrating desired multifunctional properties into self-powered sensors working under extreme environments remains challenging. Herein, highly conductive, mechanical resilient and self-healable hydrogels are developed by initial regulation of hydrogen bonding network of cellulose-Zn2+ solvation structure with further employing of borax crosslinked polyvinyl alcohol (PVA) network. Thanks to the dynamic interactions and ion chelating capability, the resultant Zn2+-Polyvinyl alcohol-Cellulose-Borax hydrogels (ZPCB) delivered remarkable stretchability (1059.5%), excellent ionic conductivity (29.4 mS cm-1), and exceptional anti-freezing property (-60 °C), self-healing capability (98.5%) and water retention ability under humid environment. With these attributes, the hydrogel assembled triboelectric nanogenerators (TENGs) demonstrated fast response (237 ms), high open circuit voltage (128 V), power density (8 W m-2) and durable output performance (5000 cycles) enabling powering small electronics and reliable signal acquisition under harsh conditions (-60 °C or placed for 45 days). Moreover, the TENG integrated with a lightweight convolutional neural network (CNN) yields 99.17% accuracy in motion-pattern recognition, encompassing clapping, jumping, slow walking, fast walking, running, and tumbling. Overall, this work offers a robust, eco-friendly, and intelligent pathway for next-generation wearable electronics and self-powered sensing systems operating in extreme environments.