| 摘要: | 非加壓式血壓量測是近年來穿戴式裝置的重要發展技術，在目前提出的技術中，以脈波傳遞速度(Pulse Wave Velocity, PWV)估測血壓之方式為普遍認為最有潛力的技術，利用心電圖(Electrocardiography, ECG)訊號的R波及光體積血容積(Photoplethysmograph, PPG)訊號之波峰間的時間差，計算出PWV並推測血壓，發展出非加壓式血壓量測技術。然而過去發展的研究，因為在參數的規劃與資料的量測上仍然有許多的不確定因素，因此導致以PWV為理論的血壓量測方法，都還在發展的階段。因此，為求數據準確信及可信度，本論文與桃園壢新醫院進行人體試驗 (Institutional Review Board, IRB)合作計畫，由專業心血管醫師篩選患有高血壓及糖尿病病患進行臨床實驗，以左手手腕作為PPG及ECG的量測點，實驗過程透過腳踏車健身器材讓受試者運動5分鐘，誘發收縮壓之提升，同時紀錄該位受試者之年齡、身高體重及病史，數據收集完畢後，利用量測到之ECG及PPG訊號計算脈波傳遞時間(Pulse Transit Time, PTT)，並利用PPG訊號特徵推算受試者之血管彈性係數，以上述之參數進行線性迴歸分析，並與經過FDA認證之生理監控儀器EDAN M50作驗證，結論符合舒張壓及收縮壓與脈波傳遞速度成正比之趨勢，在分析過程中亦得知年齡與PWV之正相關性，實驗結果，估測血壓與實際血壓平均誤差狀態為13%。在研究中發現反射式光學脈搏血氧計會受到光源亮度影響，造成血氧量測上精準度之差異，近而影響不同膚色人種之血壓估測，因此本論文針對光源變數固定完成25-65歲血壓血氧校正曲線，未來將加入自動調光機制，提高使用性及便利性。;Cuffless blood pressure measurement is an important item for development of wearable technology in recent years. Among current solutions, the pulse wave velocity (PWV)-based method has been regarded as the most potential solution for blood pressure (BP) measurement. The PWV technique measures the pulse-transit time (PTT) between the R wave of ECG and the feature of PPG wave, and uses the PPT to estimate BP. However, current PWV-based methods are still under development, because confounding factors, e.g., signal features, biomarkers, data measurement methods, etc., are still not clear which causes uncertainties in the estimated results. Therefore, to discover a better model for BP estimation using PWV-based approach, we have performed a human experiment in Landseed hospital under the supervision of Institutional Review Board (IRB), Landseed hospital, Taiwan. We have recruited 178 subjects, including normal, hypertension and diabetes patients. All subjects were asked to rest for 5 minutes (Rest State), and then were requested to perform 5 minutes bicycle stepping exercise (Exercise State). The PWV values in the Rest States and Exercise Sates were recorded to incorporate with subjects’ personal information (e.g., age, gender, height, weight) and constructed a linear model using least square regression. The estimated results were compared with EDAN M50 clinical monitoring system. In our study, we have found the PWV presented a high correlation with age and systolic/diastolic blood pressure (SBP/DBP). The estimated error is smaller than 13%. Future developments will focus on the development of algorithm for auto-adjustment of PPG source in order to make our system to adapt the measurements in people with different skin color. |
