A High‐Linearity Flexible Pressure Sensor for Pulse Physiological Signal Monitoring Enabled by Machine Learning

Abstract

Abstract With the rapid development of wearable technologies and flexible electronics, there is an increasing demand for compact, high‐performance pressure sensors capable of capturing subtle biomechanical signals. In this study, a high‐linearity flexible pressure sensor (HLFPS) entirely made of silicone is presented, whose intrinsically soft and skin‐conformal structure allows seamless integration with human skin. The sensor exhibits excellent linearity (R 2 = 0.993), high sensitivity (185 kPa −1 ), and a broad pressure range (0–125 kPa), while its soft, skin‐conformal silicone structure ensures superior wearing comfort. Human pulse waveforms are collected under multiple discrete pressure levels and subsequently recorded changes in waveforms before and after physical activity. Additionally, twelve representative pulse types are generated using a pulse generator and captured by the platform. Machine learning algorithms applied to these datasets achieve a classification accuracy of 93.83% in identifying pulse types. Furthermore, pulse signals from six human subjects are collected and classified using the same algorithm, achieving an accuracy of 94.79%, which further demonstrates the robustness and generalizability of the platform in real‐world scenarios. This fully silicone‐based, highly linear sensor—combined with real‐time onboard analysis—offers a continuous, noninvasive, and multifunctional solution for next‐generation wearable health monitoring.