ICT for Health: Optimized Cuffless Blood Pressure Measurement Using ECG, PPG and Linear Regression

Hypertension is a major risk factor for cardiovascular disease (CVD), which remains the leading cause of mortality worldwide. Continuous and accurate monitoring of blood pressure (BP) plays a fundamental role not only in early diagnosis but also in patient therapeutic management and the prevention of associated complications. However, currently used methods have significant limitations: invasive systems, while highly accurate, can only be used in hospital settings and involve risks and discomfort; noninvasive cuff-based systems, on the other hand, provide only intermittent readings and do not allow dynamic BP analysis during daily activities. To overcome these limitations, in recent years there has been growing interest in cuffless techniques that exploit physiological signals easily acquired through wearable sensors. This thesis explores a methodology for continuous and noninvasive estimation of BP based on combined electrocardiogram (ECG) and photoplethysmogram (PPG) analysis. From these signals, two key parameters are extracted: heart rate (HR) and pulse transit time (PTT). They are known in the literature for their correlation with variations in blood pressure. The work introduces several significant innovations: the use of wearable Shimmer devices to acquire signals from subjects under controlled conditions, as an alternative to data sets available online (MIMIC and others), and the design of a processing pipeline including signal preprocessing, automatic peak detection, and robust extraction of HR and PTT. Subsequently, through the application of regression techniques, a mathematical relationship was defined between these parameters and the reference BP values obtained using calibrated devices. The expected results include the development of a reliable framework capable of estimating blood pressure in real time with an error within the thresholds established by international guidelines (AAMI/ISO/ESH), the validation of the method against clinical standards and the optimization of algorithms to reduce motion artifacts. Looking ahead, such an approach could contribute to the evolution of wearable healthcare technologies, allowing continuous, discreet, and personalized BP monitoring even outside the clinical setting. This would represent an important step in the treatment of hypertension and the prevention of CVD, improving the quality of life of patients and reducing the overall social and health impact of these diseases. Therefore, an aspect of the project intended for the integration of this type of algorithm on wearable devices: the European Persimmon project and shimmer3.