A High‐Sensitivity Wearable Sensor for Precision Recognition of Human Arm Joint Movements Using CNN

ABSTRACT With the cross‐integration of flexible electronics and artificial intelligence technologies, high‐sensitivity wearable sensors have shown great potential in fields such as medical rehabilitation, human‐computer interaction, and sports science. To meet the dual requirements of high sensitivity and flexibility for wearable applications, this study proposes a novel sandwich‐structured P(VDF‐TrFE)/BTO‐OH/P(VDF‐TrFE) piezoelectric sensor (SP‐sensor) using a fused deposition modeling (FDM) process. This structure effectively combines the high piezoelectricity of BTO nanoparticles with the excellent flexibility of the P(VDF‐TrFE) polymer, overcoming the limitations of traditional single‐layer piezoelectric sensors. Experimental results demonstrate that the piezoelectric response voltage and current of the SP‐sensor are enhanced by 51.1% and 546%, respectively, compared with those of single‐layer P(VDF‐TrFE) films. With improved piezoelectric performance, the SP‐sensor achieves approximately 50% higher sensitivity than the traditional designs. It also exhibits quick response and recovery capabilities, with a response time of 8.1 ms and a recovery time of 64 ms. Additionally, it exhibits excellent fatigue resistance, with no noticeable voltage decay observed after 18,000 cycles of testing. An intelligent wrist motion recognition system, integrating a deep learning algorithm (CNN model), was developed based on this SP‐sensor, enabling real‐time classification of three types of wrist movement patterns with an identification accuracy of 97.85%. This study, through the innovation in material structure and the integration of AI algorithms, paves the way for the application of next‐generation wearable devices in human‐machine interaction and medical diagnosis.