摩擦电效应
可穿戴计算机
材料科学
高保真
忠诚
可穿戴技术
计算机科学
声学
灵敏度(控制系统)
振动
信号(编程语言)
持续监测
电阻抗
智能手表
联轴节(管道)
结构健康监测
分类器(UML)
信号处理
深度学习
状态监测
人工智能
探测理论
远程病人监护
能量(信号处理)
作者
Lirong Tang,Shanlin Yan,Xindan Hui,Yudan Liu,Yuanchao Ren,Zizhuo Wang,Zhong Lin Wang,Wenkui Dong,Fei Wu,Hengyu Guo
摘要
ABSTRACT Continuous monitoring of body‐surface vibrations via ambulatory sensing provides critical data for individual healthcare and human health mapping. However, the lack of deep structural integration along the entire skin‐to‐sensor acoustic pathway has rendered wearable mechano‐acoustic sensors incapable of balancing wearability and signal fidelity at the system level. Here, we report a flexible, skin‐interfaced triboelectric patch that addresses impedance mismatch through an internal corrugated sensing structure integrated with a membrane‐based resonant acoustic coupling interface. This architecture substantially augments mechanical‐to‐electrical transduction, encompassing the dominant energy spectra (0–1000 Hz) of physiological motions with a sensitivity of 2062.66 mV·Pa − 1 ·cm − 2 at 168 Hz with a low detection threshold of 45 dB. Then, a bispectrogram classifier based on the Cardiechema database is constructed, tuned, and evaluated via 5 × 3 cross‐validation, achieving a mean accuracy of 96.71% and a macro F1‐score of 96.72%. By synergizing high‐fidelity sensing with machine learning, this robust multimodal framework establishes a general approach for real‐time diagnostics and personalized healthcare, offering profound insights into complex physiological signatures from cardiac pathologies to intricate sleep architectures.
科研通智能强力驱动
Strongly Powered by AbleSci AI