Intelligent Gait Analysis System Enabled by Liquid Metal-Embedded Sponge Triboelectric Sensor Arrays

Gait dynamics are pivotal biomarkers for early disease prediction and human health assessment. In this study, we propose an intelligent monitoring system that integrates flexible PDMS/liquid metal sponge triboelectric nanogenerator (PLMFT) arrays with convolutional neural networks (CNNs), enabling comfortable, long-term gait monitoring. The PLMFT device features a porous matrix infiltrated with liquid metal, which gives the sensing unit excellent mechanical flexibility, high electrical output, and robust mechanical stability over 3000 compression-release cycles; based on this, an insole-type monitoring system is constructed, which integrates five sensing units in a flexible substrate and combines both breathability and wearable comfort. A convolutional neural network (CNN) is used to analyze the collected gait signals, and the recognition accuracy is as high as 98.95%. This work presents a high-precision and lightweight solution for wearable health monitoring, offering greater potential for application in gait abnormality detection, motor function assessment, and disease prediction.