Research On Pulse Signal Acquisition System Development And Processing Method

Pulse is the pulse of arteries caused by the blood circulation system,which can reflect part of the physiological and pathological information of the human body.Traditional Chinese medicine pulse diagnosis is to obtain the pulse information of the human body by touching and pressing the wrists of the human body,so as to complete the disease diagnosis.However,due to the lack of unified diagnosis and treatment standards,pulse diagnosis results are greatly affected by doctors’ subjective factors.Therefore,the development of pulse signal acquisition system and pulse signal processing is of great significance to the digitization and objectification of pulse diagnosis.This paper developed a wristband pulse sampling system based on the Cunkou "Three Portions and Nine Pulse-takings" theory of traditional Chinese medicine,which uses three independent wristbands to flexibly locate the Cun,Guan and Chi of different Cunkou parts of human body.The pressure of the wrist band is changed by the air pump to simulate the way of taking the pulse of the finger in pulse diagnosis of traditional Chinese medicine.The PVDF flexible sensor was used to collect pulse image,and the sensor acquisition structure with high sensitivity was designed through simulation.According to the characteristics of the output signal of the sensor,the preprocessing circuit and digital acquisition circuit are designed to complete the filtering and de-noising of pulse signal,A/D conversion and USB transmission.The upper computer is developed on the platform of Lab VIEW,and the functions such as instruction sending,pressure control,pulse data acquisition,data storage and result display are realized.After testing,the pulse collecting system designed in this paper meets the requirement of Cunkou "Three Portions and Nine Pulse-takings" in traditional Chinese medicine.In this paper,an improved EMD algorithm is proposed by combining single-period contrast with EMD algorithm,and the pulse signal is filtered to remove the false baseline in the pulse signal effectively.Nektar 1D,a new numerical model of pulse wave propagation,was used to generate pulse signals of the radial artery of the wrist in 384 cases of 6 age groups using cardiovascular parameters of different ages.The generated pulse signals were extracted in time and frequency domain and dimensional reduction were performed.Combined with the pulse signal samples of different ages and the feature information after dimensionality reduction,the decision tree algorithm and SOM algorithm were used for classification and recognition training,and the training results were obtained.Pulse signals from six parts of the left and right hands of 18 students were collected under the pressure of "Fu,Zhong and Chen" for classification and recognition.The results show that the SOM algorithm has better accuracy than the decision tree algorithm and is suitable for the study of pulse signal pattern recognition.