Wearable Application-oriented ECG Signal Processing Method And Software Implementation

With the improvement of people's living standards and the quickening pace of life,the morbidity of heart disease is rising rapidly and it becomes one of the major factors that threaten human's health.Heart attack is rapid and its attack time is uncertain,it needs long lasting monitoring of human's heart activity.Electrocardiogram is a graph that records the change of electrical activity generated by each cardiac cycle of the heart.It is the main basis for diagnosing heart disease,and has the advantages of briefness,reliable diagnosis,and no harm to the patient.Using wearable computing technology to integrate fabric electrodes,analog front ends,processors,and communication circuits in chest strap or other forms of clothing to enable long lasting ECG monitoring.The main work of this paper is designs related ECG signal processing methods for such wearable ECG monitoring system and completes software implementation.The system can collect,store and process ECG signals in real time,extract features,calculate ECG feature parameters,and detect abnormal ECG signal segments in real time according to preset or user-defined strategies.The research contents of this paper are summarized as follows:(1)Combine the related technical theories of ECG signal processing and analysis to do requirement analysis on the software system,and complete the overall design of it.The software system can be divided into modules according to its functions,including: ECG signal data acquisition and storage,preprocessing,feature extraction,ECG signal analysis,visualization and user interaction.The design requirements for each functional module has been stated.(2)Design each functional module of the software system separately,including: the process of ECG signal transmission,the process of ECG data package parsing,and the process of ECG data storage implementation;the implementation process of using wavelet transform to remove baseline drift and using a method by combining empirical mode decomposition and principal component analysis to eliminate EMG interference,and use the ECG data in MIT-BIH database to simulate the above two algorithms;the implementation process of ECG feature extraction using differential threshold method and wavelet transform;the implementation process of using logic branch judgement method to detect the abnormal ECG signal segments;the overall layout of the software system,real-time display of ECG waveforms and the components that implement user interaction.(3)Coding to implement the whole software system by using Python programming language,do functional test on the software system,and verify that each functional module of software system meets the design requirements.The signal-to-noise ratio,mean square error,heart rate sensitivity and variability are used as indicators to evaluate the performance of the software system using ECG data acquired by experiments.The calculation results show that in the case of resting and slow walking,the effect of ECG signal preprocessing is good,and the R-wave location achieves high accuracy.Using the data from the MIT-BIH Arrhythmia database to test abnormal ECG detection algorithms,the accuracy of the anomaly detection is 93.3%.Based on the research,the ECG signal analysis of software system proposes a user-defined diagnostic function and adds a new kind of detected abnormal ECG signal by configuring a diagnostic rule.