Non-invasive cuff free blood pressure and heart rate measurement from photoplethysmography (PPG) signal using machine learning

Measuring the blood pressure (BP) and heart rate (HR) is essential in order to monitor the physiological vital parameters of patients admitted in Intensive care unit (ICU). Development of precise noninvasive measurement devices are encouraged for better healthcare facilities. Noninvasive methods are preferred for painless and patient friendly measurements. The existing cuff based measuring devices exerts pressure in arms which irritate the patients when intravenous solutions are administered through hand nerves. To overcome the inconvenience and continuous BP measurements, a novel Photoplethysmography (PPG) based BP, heart rate (HR) monitoring measuring device is proposed. The proposed algorithm uses Principal Component Analysis (PCA) to get the required features from the PPG signal five different machine learning (ML) algorithms have been analyzed for the prediction of blood pressure and heart rate. Support Vector Regression (SVR) algorithm outperforms the other ML algorithms. The proposed algorithm is implemented in hardware using a reflectance Pulse sensor and Raspberry Pi microcontroller. The hardware results are compared with those of commercially available devices, indicating that the device serves as a noninvasive tool for measuring blood pressure and heart rate with an accuracy of approximately 98%.