A Flexible Smart Insole Enabling Real-Time Gait Visualization and Phase Recognition

The widespread adoption of wearable plantar pressure insole systems has long been hindered by limitations such as low spatial resolution, insufficient data processing intelligence, and high sensor crosstalk. Here, we present an intelligent insole system featuring a flexible, high-density piezoresistive sensor array with a sensor density of 1.1 sensors/cm2 , a wide dynamic range (0–300kPa), and high durability (over 100,000 compression cycles). Each insole integrates 238 uniformly distributed sensing units per foot in an orthogonal matrix, coupled with a wireless, low-power acquisition module for high-frequency, real-time plantar pressure monitoring in natural settings. The system enables comprehensive visualization and analysis of plantar pressure distribution and gait patterns during various activities, such as walking, jumping, in-toeing, and simulated injury. It can intuitively assess foot loading and gait characteristics, facilitating applications in sports biomechanics, rehabilitation, and early disease detection. In addition, we introduce a Plantar Pressure-based Gait Network (PPGaitNet) for automated gait phase recognition, enabling quantitative analysis of locomotion and providing a data-driven foundation for embodied intelligence, such as exoskeleton gait following and adaptive rehabilitation. Experimental results confirm that the system delivers robust, high-fidelity plantar pressure data and reliable gait phase segmentation with a classification accuracy of up to 93%, highlighting its promise for biomechanical analysis, clinical evaluation, and intelligent human–machine interaction.