Research On Monitoring Algorithm Of Respiratory System Diseases Using Portable Device ECG

With the development of society,people pay more and more attention to their physical and mental health.However,some respiratory diseases,especially chronic obstructive pulmonary disease(COPD),are not obvious in the early stage,thus causing people to ignore their harm.COPD is difficult to detect at an early stage because of its slow progression,and its diagnosis depends on multiple conditions.By the time it progresses to dyspnea,it is already in an irreversible stage and requires long-term medication.Early treatment can reduce the impact of COPD on health.Studies have found that there are differences in ECG between patients with COPD and normal people,and ECG is widely used in medical institutions,so it is necessary to analyze ECG to determine whether patients with COPD.Wearable technology has developed rapidly in recent years,and wearable devices in the medical field are also booming.Wearable ECG monitoring devices can solve the long-term physiological data monitoring,and remote monitoring system to achieve non-contact diagnosis,reducing the burden on individuals and hospitals.ECG signals collected by wearable ECG monitoring devices are prone to be affected by ambient noise.In this study,several common noises are processed accordingly.A moving window average is constructed based on a principle similar to convolution,and a variety of noises are filtered through the design of moving average window size.In this study,we propose a QRS group monitoring method based on the adaptive threshold method.The method includes the detection of the morphology of R wave in ECG,and the feasibility of the algorithm is verified by using the data of MGH/MF.The accuracy of R wave localization was verified by MIT-BIH arrhythmia database,which was 97.63%.On the basis of R wave detection,Q wave and S wave are searched by the maximum value method and slope limit.After QRS localization completed,P and T wave localization are based on morphological feature detection of P and T wave.Thirteen features,including the amplitude and interval of each characteristic wave,were extracted by ECG correlation point localization.Then,the hypothesis-deductive method was used to assign the same label "1" to all the features extracted from the ECG data of COPD patients,and "0" to the features extracted from the ECG data of the patients without COPD disease.From four different types of classifiers including Bayesian classifier and SVM,the decision tree and clustering characteristics of 13 kinds of repeated experiments,a series of analysis and analyze the results of the experiment,so as to find out whether the diagnosis way.It is using the way of clustering in COPD patients’ ECG characteristics of single binary classification,then the category with a large number in the model is given a label "0",while another kind is "1".By analyzing predictive classification results to determine whether patients with COPD or not.Finally,this study designed a wearable monitoring system and built an ECG monitoring APP based on Android platform to realize Bluetooth connection,data transmission and data storage with the wearable device.This wearable monitoring system is convenient to use,and the Android APP has simple functions,simple operation and practical.