Research And Application Of Arrhythmia Detection In Cardiac Exercise Rehabilitation System

With the development of technology and the improvement of medical care,increasingly attention paid on the cardiac exercise rehabilitation.The cardiac exercise rehabilitation system can provide online real-time monitoring as well as scientific and effective management of the rehabilitation process of patients with heart disease,and reduce the risk of patients during exercise rehabilitation.This paper aims at the automatic detection of arrhythmia in the online real-time monitoring of cardiac exercise rehabilitation system,and studies the data analysis and processing methods of ECG and classification methods of arrhythmia.Besides,this paper made a specific design and development in combination with the research and development of cardiac exercise rehabilitation system.The main work of this paper is as follows:(1)This paper analyzed the research status of the original ECG noise processing,waveform feature extraction and arrhythmia classification at home and abroad through literature review.(2)The paper studied the wavelet transform de-nosing algorithm and determined the number of wavelet decomposition layers by comparing the appearance of the ECG signal’s global feature under different scale decomposed.Then paper compared the results of removing high frequency noise and base line drift using four kinds of wavelet on the original ECG.Then paper chose the wavelet function with a higher SNR value to construct the wavelet basis to finish the preprocessing of original ECG signal and finally got a relatively pure signal.(3)This paper identified and located the characteristic waveforms of QRS,P,and T waves of ECG signals using PT,slope-based,relative position-based algorithms,and extracted key features such as RR interval and R amplitude from ECG signals.After detecting five main waveforms of ECG signal,this paper combined the various annotation information added by experts of ECG records in the MIT-BIH database with the patient’s age,gender and medical status to construct a feature set.Then the paper extracted the target four types of heart rhythm from the database to study and classify.After that,the paper used support vector machine and decision tree algorithm to classify four types of heart rhythm and comparing the F2 value and ROC curve of the two classification results,it is determined that the decision tree have a better performance on the classification of four types of heart rhythm especially on three arrhythmia types.(4)Based on the above work,this paper designed and implemented a mobile client for the cardiac exercise rehabilitation system.So that the mobile monitoring client can not only acquire real-time ECG data of the cardiac exercise rehabilitation patient through the wireless AP network,but also can display the ECG waveform.In addition,this mobile client can identify and classify arrhythmia.Therefore,it can make a warning of arrhythmia occurring in the process of patient rehabilitation exercise and effectively reduce the risk of exercise rehabilitation for patients.