Research On BCG Signal Acquisition System And Adaptive Processing Method Based On Piezoelectric Cable

The rapidly developing social environment has fundamentally changed people's lifestyles,and more and more people are plagued by sub-health problems.Cardiovascular disease is the most typical category of sub-health problems,and both its morbidity and mortality are high.However,the electrocardiographs and other equipment used to diagnose cardiovascular disease in modern medicine have not been widely applied in homes for various reasons.To some extent,patients have delayed the optimal treatment period.In view of the lack of human health monitoring equipment at home,a physiological signal detection system based on PVDF piezoelectric cable sensors is designed and researched in this paper.The system has sensors built into the mattress and enables patients to grasp their health information in time by collecting and analyzing ballistocardiography(BCG)signals.The main research work of this paper is as follows:Firstly,according to the characteristics of PVDF piezoelectric cable and BCG signal,an S-shaped piezoelectric cable sensor and its corresponding hardware circuit were designed,which mainly include a sensor module,a signal conditioning module,a microprocessor module,a communication module,and a power module.Then,the research on the adaptive method of heart rate extraction was carried out,and a new adaptive algorithm combining the differential threshold peak-seeking algorithm and the unsupervised learning clustering algorithm was proposed and implemented.By matching the obtained heartbeat template with the newly acquired BCG signal,the heart rate information can be obtained in real time.At the same time,human body sleep was staged through body motion signals combined with heart rate information,and the recognition of arousal,eye movement,light sleep and deep sleep during sleep was realized.Secondly,in order to make the device more intelligent,a systematic study was performed on the heart rate abnormality classification(HRV)function.According to the correlation between ECG signal and BCG signal,the feasibility of HRV analysis using BCG signal was demonstrated.In view of the lack of heart rate abnormal BCG data samples,ECG heart rate abnormal data samples in the MIT-BIH library were used to train and study BP,RF,and SVM HRV classification models.The performance of the three models was evaluated through the confusion matrix and other related indicators,and the BP neural networkclassification model was found to be optimal.Finally,in order to make the real-time physiological information detection system more humane and more convenient to use,the upper and lower computer system software was designed and completed.The?C/OS-IIIoperating system was transplanted in the lower computer to realize multi-task management and division.The upper computer designed a human-computer interaction interface based on QT software.After system experiments and tests,the BCG signal detection system based on piezoelectric cables developed in this paper can realize the real-time communication function of the host computer and the lower computer through the serial port module and the Bluetooth module,and can collect and process the patient's body seismic signal to a computer or mobile cloud.The heart rate measurement error is about 2.65 times/minute and the accuracy reaches96.77%.The constructed BP neural network classification model can basically classify the abnormality of BCG signal heart rate,with an accuracy of about 74%,and achieves the expected function.The research results of this paper have certain reference value for the development and design of related instruments based on the BCG signal measurement principle.