A new system for cuffless blood pressure measurement

High blood pressure (HBP) increases the risk of cardiovascular, brain, and renal diseases which can lead to severe illness and death. Due to the variability of BP values at different times during the day, continuous BP measurement would allow more effective early diagnosis of elevated blood pressure, and aid in the management of treatment. Various techniques have been proposed for cuffless BP measurement, often relying on the estimation of pulse transit time (PTT) to infer BP. However, challenges such as computational complexity and limited availability of BP monitoring sensors hinder the reliable and widespread implementation of continuous BP measurement. The proposed wristwatch-based system uses speech sounds together with photoplethysmogram (PPG) and electrocardiogram (ECG) signals, and uses a mathematical model which estimates the systolic BP (SBP) and diastolic BP (DBP) from these inputs. We have obtained Pearson correlation coefficient (PCC)s of 0.6094 for systolic BP and 0.6001 for diastolic BP in these experiments. The mean average errors (MAE) obtained are 8.06 mmHg for systolic BP and 7.48 mmHg for diastolic BP, respectively. This study provides meaningful results based on experimental outputs, demonstrating the feasibility of using speech signals in conjunction with PPG and ECG signals for BP estimation. The simplicity of our mathematical model enables efficient and practical implementation in real-time BP monitoring. As a pilot study, our work establishes a foundation for further research, emphasizing the need for enhancing accuracy and reliability, expanding the dataset, and exploring additional physiological parameters to advance continuous BP estimation. Our findings advance the development of cuffless BP monitoring, and address the computational complexity and real-time processing requirements. Future studies should build upon our work to refine and optimize the proposed approach, ultimately leading to improved continuous BP estimation in diverse populations.