Research And Application Of Real-time Data Anomaly Detection System For Motion ECG

During exercise,some people may experience physical discomfort,and severe death may occur.Therefore,when exercising,people should pay more attention to their heart condition and adjust the amount of exercise in time,so as to avoid irreparable consequences.Electrocardiogram can describe the human heart's electrical activity,and it is also an important basis for doctors to diagnose and treat heart disease.It has practical application value for real-time ECG monitoring of people in motion.In this paper,ECG anomaly detection methods are studied.ECG signal analysis is mainly divided into four processes,the steps are:data preprocessing,QRS complex detection,feature extraction and ECG anomaly recognition.According to the existing intelligent algorithm,in order to ensure the de-noising effect of ECG signal and the real-time performance of the algorithm,the filter used in this paper is a band-pass filter composed of cascaded low-pass and high-pass filters.This filter can well remove the noise in the motion ECG signal.The Pan-Tompkins algorithm is used to detect the QRS complex in real time.In the analysis of ECG data,extracting accurate ECG feature is the key link in ECG data analysis.This paper proposes a method of ECG feature extraction based on wavelet packet decomposition and principal component analysis.In order to improve the accuracy of ECG anomaly recognition,genetic algorithm is used to optimize the parameters C(penalty parameter)and g(kernel function parameter)of support vector machine,and a GA-SVM classifier is designed.In this paper,the treadmill exercise ECG data provided by Beijing Mai bang Photoelectric Instrument Co.,Ltd.and the four types of ECG data of the international arrhythmia database MIT-BIH:normal heartbeat,ventricular premature beat,left bundle branch block,right bundle branch block are used for experiments.The classification effect of the test set is good,and the accuracy is better than other classified prediction models.The experimental results show that the proposed feature extraction algorithm is stable and effective,and the performance of the classifier is improved by optimizing the parameters of support vector machine through genetic algorithm.Therefore,the ECG data anomaly detection method used in this paper can effectively identify arrhythmias,which is of great significance for the diagnosis and timely treatment of cardiac diseases.Finally,this paper develops a motion ECG anomaly detection system based on Android.The App software is used in conjunction with the corresponding ECG acquisition equipment.The system can complete the real-time transmission of ECG data and analyze data on the mobile phone.In addition,the system also has the function of displaying the user's motion trajectory and locating the user's position in real time.Once the abnormal ECG of the user is found,the system can alarm in time.After testing,the performance and function of the software run well.