• 전기학회논문지
• /
• 제57권6호
• /
• pp.1095-1101
• /
• 2008

The desirable method to diagnose abnormal blood pressure is to measure and manage blood pressure continuously and regularly. However, the sphygmomanometers that are based on a cuff have faults in that they can not measure the blood pressure continuously and they cause an unpleasant feeling. Therefore, it is essential to develop a new measuring method that causes no pain and that can obtain blood pressure continuously without any unpleasant feeling. Thus, we propose here a regression method to estimate the systolic blood pressure by using the PTTL(pulse transit time on leg) with some body parameters which are chosen from the relational analysis with systolic blood pressure. The data we use to make the regression model were obtained in triplicate from each of 50 males who were from 18 to 35 years. And we made estimation experiments of blood pressure on 10 males who did not take part in the making the regression model. According to the results, the proposed method showed a mean error of 4.00 mmHg and the standard variance was 2.45 mmHg. When we comparing the results of the proposed method with the rule of American National Standards Institute of the Association of the Advancement of Medical Instruments(ANSI/AAMI), the results satisfied the rule of a mean error less than 5 mmHg and a standard variance less than 8 mmHg. Therefore we were able to validate the usefulness of the proposed method.

• 한국산업정보학회논문지
• /
• 제25권3호
• /
• pp.1-10
• /
• 2020

본 논문에서는 시계열 심전도 (Electrocardiogram: ECG) 및 광전용맥파 측정센서 (Photoplethysmography: PPG)을 이용하여 혈압을 추정하는 알고리즘을 제안한다. 혈압 (Blood pressure: BP)을 추정하기 위해 주기적 입력 신호를 생성하고 차동 및 임계값 방법에 따라 잡음을 제거한 다음 합성곱 신경망 알고리즘을 기반으로 하여 수축기 혈압과 이완기 혈압을 예측한다. 본 논문에서 사용된 데이터는 MIMIC 데이터베이스에서 총 3.1GB의 49명의 환자 데이터를 사용하였다. 실험결과 수축기 혈압의 평균 제곱근 오차는 5.80mmHg, 이완기 혈압의 예측 오차는 2.78mmHg을 나타내었다. 또한, 영국 고혈압 협회가 제안한 혈압계 평가 방법을 적용하였을 때, 최고 성능인 등급 A를 만족함을 확인할 수 있었다.