Wearable textile sensors are key enablers for motion monitoring, e-health, and human digital twins. However, most existing textile pressure sensors suffer from limited operating ranges and depend on dense sensor arrays and high sampling rates, which constrain their ability to capture full-spectrum plantar pressures while maintaining communication and energy efficiency. Here, we present an intelligent dual-mode textile sensing system that is flexible, soft, and seamlessly integrable into garments, designed to mimic the hybrid mechanosensation of human skin. An ultra-broad linear pressure sensing range of up to 800 kPa is achieved, making the sensor well suited for foot applications involving large and rapidly varying loads. By embedding a strategically optimized set of sensing nodes into conventional socks, the system enables reliable spatiotemporal mapping of plantar pressure distributions without dense sensor arrays, supporting real-time detection of gait abnormalities. Furthermore, plantar pressure features are processed using a one-dimensional convolutional neural network to classify 14 distinct body postures. Owing to the sensor’s high signal-to-noise ratio and linear response, a posture recognition accuracy of 98.45% is maintained even when the sampling rate is reduced by half. By jointly minimizing sensor density and sampling frequency without sacrificing performance, the proposed system significantly reduces communication bandwidth and power consumption. Overall, this intelligent textile sensing platform offers a scalable and communication-efficient solution for spatiotemporal plantar pressure monitoring, enabling continuous gait assessment, healthcare monitoring, and serving as a physical–virtual interface toward real-time human motion digital twins in everyday environments.
Jiang et al. (Tue,) studied this question.