具心電訊號擷取及判讀之手機照護系統

The Health Care System of Mobile Phone with Detecting and Analyzing of ECG

鄭淳銘(Chun-Ming Cheng)

指導教授 ： 翁清松

中原大學/工學院/生物醫學工程研究所/碩士(2008年)

https://doi.org/10.6840/cycu200901308

摘要

由於醫療科技的進步，心電圖已是心臟疾病診斷及治療最基本的檢查項目之一。有些心臟疾病如心室顫動(ventricular fibrillation)等，病發時若無及時實行救護，往往為造成病患猝死之原因。一般來說，需要長期監測的病患，都以配帶霍特(Holter)心電圖記錄器，因此限制了病患的活動力。為改善病患生活品質及保障其生命安全，本研究發展一結合手機之心臟疾病辨識系統，並擁有當生命受到威脅時可緊急呼救的救援系統。
本研究將一般擷取第一導程心電圖方式從三個電極改良為兩個電極，且使用乾式電極，擷取心電圖的電路包含了放大器、濾波器等，且使用訊號整合控制處理器MSP430將擷取後的心電訊號做類比/數位轉換，透過藍芽模組使用串列傳輸(UART)訊號傳輸至PDA手機後，使用環狀結構的數位濾波器將訊號再次濾波及使用Tompkins演算法擷取心電圖中最重要的R波後，將P、Q、S、T波的位置及振幅各別取出運算出與其相關的12項特徵作為判別病症之參數，最後透過類神經網路訓練學習並求出判別法則。在硬體方面順利以雙電極方式將心電訊號擷取並通過各項濾波放大後訊號良好，且通過藍芽傳輸後並無封包遺失之現象。在臨床實驗上，選用了正常人（訓練組：16人、判別組：12人）、患有冠心病（訓練組：8人、判別組：12人）及裝置心律調節器（訓練組：5人、判別組：7人）的病患，將其心電訊號取出及輸入類神經網路訓練後，在判別組病症上也達到了平均為97.2%的準確率。訊號亦同時記錄於記憶體內，如果病患發生緊急狀況，則可透過簡訊服務(short message services,SMS)將求救訊號自動或手動發送至救護中心。
本研究提出整合GPS定位、藍芽傳輸、自動心電圖病症辨識與簡訊服務，亦可針對高齡化的人口 (弱勢獨居老人及失智老人)提供高品質照護與協尋能力。本研究最終目標為發展具實用性之生理監控與緊急救護系統，可供醫護人員或家屬有效掌握病患(或老人)的動向及健康資訊。

關鍵字

MSP430 ； 心電圖 ； 冠心病 ； 類神經網路

並列摘要

As a result of advances in medical technology, ECG is the one of basic examination on heart disease. Some heart diseases such as ventricular fibrillation are the major causes of sudden death. In general, the activities of patients monitored by Holter’s recorder are restricted by the recorder. In order to improve patients quality of life and to ensure their safety, we combined a cellular phone with a heart disease recognition system and emergency rescue system.
In this study, we modified the number of electrodes from three to two electrodes of Lead I in ECG and used dry electrodes. The parts of the hardware include amplifiers, filters and so on. The MCU(MSP430) was used to receive the signal and to do Analog/Digital conversion and transfer to PDA phone through Bluetooth module with Universal Asynchronous Receiver/Transmitter (UART). The signal was filtered through digital filter again and made use of Tompkins algorithms to retrieve the R-wave of ECG in PDA and MSP430. By positioning of R-wave, we could determine the amplitudes and times of P, Q, S and T waves at the same part of ECG, and derive 12 parameters for the use in the rule of neural network through these points. Regarding to hardware, the dual-electrodes were used to capture the ECG signal successfully, and the signal-to-noise ratio of the signal waveform through the filter was high. The packet loss phenomenon didn't appear during transmission in Bluetooth. In clinical, the normal adults without any cardiac diseases (train: 16 persons, discriminant: 12 persons) were recluded. The patients (train: 8 persons, discriminant: 12 persons) that suffers from coronary artery disease and the patients (train: 5 persons, discriminant: 7 persons) with pacemaker were studied. The results showed that the parameters of their ECG for the diagnosis of the diseases in neural network achieved an average 97.2% accurately. The ECG was also recorded in the memory of PDA at the same time. The patient could send the message automatically or manually to the rescue center through SMS.
The device integrated GPS, Bluetooth transmission, automatic disease identification from ECG and message services in this study. It also provided high-quality care and search function for the old man (elders who live alone and dementia elders). The ultimate purpose of this study have relevance for the development of physiological monitoring and emergency rescue system and support the situation and health information for health care workers or family members of patients.

並列關鍵字

MSP430 ； ECG ； coronary heart disease ； neural network

參考文獻

【25】童淑芬，類神經網路在心電圖分類之應用,中原大學醫學工程

華民國體育學會體育學報， 11: 131-150，1989

【2】J. B. Conti, R. M. Mills, D. A. Woodard and A. B. Curtis, 「QT dispersion is a marker for life-threatening ventricular arrhythmias after atrioventricular nodal ablation using radio frequency energy,」 The American Journal of cardiology, 97: 1412, 1997.

【3】J. P. Singh, P. Sleight, A. Kardos and G. Hart, 「QT interval dynamics and heart rate variability preceding a case of cardiac arrest,」 Heart, 77: 375-377, 1997.

prolongation is associated with inducible ventricular tachycardia after repair of tetralogy of fallot,」 The American Journal of Cardiology, 80: 160-163, 1997.