A Force Sensor-Based Approach for Continuous Blood Pressure Monitoring on Wearable Devices

Continuous monitoring of blood pressure holds significant importance in preventing cardiovascular diseases. Non-Invasive Blood Pressure (NIBP) techniques are particularly valuable for this purpose as they offer a non-invasive alternative to traditional cuff-based methods, providing users with a more comfortable and user-friendly experience. In this paper we propose using Forcecardiography (FCG) and Edge AI techniques to advance real-time, all-day cuffless blood pressure monitoring. The use of FCG signal from a force sensing resistor (FSR) is validated by comparing it against established Photoplethysmography (PPG) methods. One key advantage of FCG over PPG is its reduced energy consumption since it does not require LEDs, enabling low-power continuous monitoring. By integrating Edge AI processing, we aim to enable efficient and accurate monitoring beyond traditional approaches. Neural Networks (NNs) developed in this study align well with IEEE and AAMI standards, producing results in terms of Mean Absolute Deviation (MAD), Mean Error (ME), and Standard Deviation (SD) that meet established criteria. A comparative analysis of mean-blood pressure (MBP) estimation using both PPG and FCG signals reveals promising results. Furthermore, real-time validation using a 32-bit microcontroller demonstrates the practical feasibility of the proposed approach.