Intelligent cardiovascular disease diagnosis system combined piezoelectric nanogenerator based on 2D Bi2O2Se with deep learning technique

Abstract

Two-dimensional (2D) piezoelectric nanogenerators (PENGs) have great potential in capturing weak physiological signals on the body surface due to their superior flexibility and high sensitivity. However, there are few reports on utilizing 2D PENG to capture physiological signals for identifying human health status, and to conduct clinical trials. Here, we reported an intelligent cardiovascular disease diagnosis system that integrates a self-powered pulse sensor based on 2D Bi₂O₂Se PENG and deep learning (DL) technology to accurately identify nine common cardiovascular diseases. The significant piezoelectricity of Bi₂O₂Se nanosheets prepared by chemical vapor deposition was experimentally validated via piezoresponse force microscopy. The open-circuit voltage (V_OC) and short-circuit current (I_SC) of the 2D PENG can reach 60 mV and 2 nA under 0.6 % strain, respectively. The 2D PENG attached to the skin can sense the three characteristic peaks of pulse signals in different body statuses, and then extract some predictive indicators reflecting cardiovascular diseases. Clinical trials show that the intelligent cardiovascular disease diagnosis system can accurately identify nine common diseases with a recognition accuracy of 93.75 %. Our study indicates that 2D PENG has an enormous potential for long-term non-invasive health monitoring, and provides a new strategy for early diagnosis of cardiovascular diseases.