Wearable ECG Signal Quality Comprehensive Evaluation And Rhythm Analysis System

Heart disease is a disease of serious harm to people's health. There are millions of people died of cardiovascular disease in the world every year, accounting for one third of the total deaths, according to statistics. In China, the number of death from cardiovascular disease accounts for about 44% of the total number of death. Therefore, the study of cardiovascular disease prevention, diagnosis and treatment has become an important issue for the medical profession. Wearable technology is to embed the ECG Acquisition System in clothing without affecting the wearing comfort so that the system can obtain ECG data in natural state. It is an effective method to monitor the ECG signals in real time.However, the human bodies are often in the state of motion and the dynamic ECG signals are interrupted greatly by noise, which has brought great difficulties to the latter analysis of the ECG, and can easily lead a miscarriage of justice to automatic guardianship, affect the timely diagnosis and treatment of patients. Although the signal quality assessment techniques can effectively prevent the movement ECG interference impact, they evaluated the ECG obtained in static condition.Abnormal ECG rhythm analysis is the core of the ECG monitoring system that can automatically diagnose the patient's condition and reduce the workload of doctors, as well as timely diagnosis and treatment of the patient. Difficulty in feature value detection as well as difficulty detection between abnormal waves lead by ECG signal itself irregular and shape varied, and some classification method is not perfect until now, so there are many basic problems are still unresolved until far.This paper focused on motion signal quality assessment and abnormal rhythm analysis, design and research on wearable ECG monitoring system.Firstly, this paper gives an overview of the automatic analysis of ECG and current research status of ECG signal processing technology, in-depth study of the ECG waveform characteristics, arrhythmia characteristics and performance, introduces some common arrhythmia database, lay the theoretical foundation for the next research.Secondly, according to design principles and requirements of wearable ECG monitoring system, we design architecture of the whole system and give out the specific design ideas and design methods of each modules.Thirdly, with a comprehensive analysis to R-wave detection match degree, power spectral density ratio and kurtosis,3 indexes for electrocardiogram (ECG) signal quality, this paper established a quality assessment model for a wearable ECG-signal based on fuzzy composite assessment. The model can assess the the signal quality and play as a foundation for next intelligent diagnosis.Next, this paper gives two kinds of abnormal heart rhythm classification method including Learning Vector Quantization (LVQ) neural network and branch of logic determine method, give a waveform classification based on signal quality index, that is choose ECG diagnostic classification according to signal quality index. It can meet real-time and precision in this way.The design of the ECG signal pre-processing, characteristic value detection and classification diagnosis in this paper is mainly implemented on the platform of MATLAB, and verified by international universal MIT-BIH arrhythmia standard database. At last, we obtained a good results by using this algorithm to analysis the data collected by this wearable ECG system.Finally, this paper focuses on design and research of the system monitoring software, give specific design steps, including the design of each functional module, propose an overall design of a wearable ECG monitoring system.The achievements will be the basis for development of ECG monitoring products for people in community-oriented, hospitals, geracomium and so on.