Study Of Heart Sound Detection System Based On PVDF Piezoelectric Thin Film Sensor

As the most import physiological indicator,heart sound signal contains a lot of valuable information of heart.Requiring and recognition of heart sound signal have great significance on the diagnosis of organic heart disease in medicine.However,limited to the frequency range of human hearing,traditional methods are difficult to teach and techniques of heart auscultation are difficult to master.With the introduction of the concept of"smart medical treatment",heart sound signal detection and identification,which with the help of medical equipment,computer and digital signal processing techniques to help people learn more about the health of their heart in timely and quickly,has become the research hot spots also difficulties in recent years.PVDF piezoelectric film,as a polymer material,has high flexibility and good piezoelectricity and thus it is very suitable for the manufacture of piezoelectric sensors detect heart sound signals.Firstly,we developed a desktop electrostatic spinning experimental device which bears advantage such as low cost,green,spinning parameters controllable.Based on this experimental platform,a good performance PVDF piezoelectric film with average fiber diameter of 0.24um,piezoelectric constant d₃₃3 of 21pC/N,relative permittivity of 11.8 and dielectric loss tangent of 0.027 was prepared by electrospinning technique.Then,from the point of portability,the design utilized a self-made PVDF piezoelectric film to develop a PVDF piezoelectric heart sound sensor.It has perfect characteristics such as higher performance stability,flexibility and sensitivity of not less than 1Vmm based on"sandwich"structure.We have designed a heart sound signal acquisition system consists of PVDF piezoelectric heart sound sensor and heart sound signal conditioning hardware circuits part and heart sound data collection part.Firstly,according to the characteristics of heart sound signal,the designing and simulation of the buffer circuit,filter circuit and amplifier circuit are done.Then,we successfully realized heart sound signal acquisition,storage and visualization based on LabVIEW graphical programming environment.Considering the introduction of noise when collecting heart sound signals is inevitable,the heart sound signals is denoised firstly.To denoise the heart sound signal,based on the wavelet transform threshold denoising principle,this paper presents an improved bivariate adjustable threshold function.Compared with several other denoising methods,the experimental results show that the mean square error and signal-to-noise ratio of the denoised heartbeat signal proposed by this paper are better than those of the common threshold function.According to the features of heart sound signal,MFCC feature vector extraction algorithm is used to extract its features and SVM algorithm is used to recognize the heart sound signals.The heart sound dataset used in this paper included both 3500 cases of normal and abnormal heart sound signals provided by the public heart sound database,abnormal heart sounds including coronary heart disease,aortic stenosis,aortic valve insufficiency,etc.;also includes the PVDF piezoelectric sensors to collect 12 cases of normal heart sound signals,4 cases of cardiovascular disease in patients with abnormal heart sounds.The recognition rate to positive and abnormal heart sound signals given by our trained heart sound recognition model is as high as 94.60%.The experimental results show that the proposed PVDF piezoelectric thin film sensor along with the proposed method of classification of positive and abnormal heart sound signals have high practical value,which lays the foundation for the diagnosis and clinical application of cardiovascular diseases.