AI‐Based Non‐Invasive Continuous Blood Pressure Monitoring and Estimation System

ABSTRACT This project explores the development and feasibility of a non‐invasive biosensor system designed to capture cardiac signals and evaluate the validity of estimating blood pressure over a 24‐hour period using a single‐lead electrocardiogram (ECG). The system integrates an Apple Watch, a laboratory‐built digital stethoscope, and a consumer‐grade blood pressure cuff to collect cardiac signals that are processed through a cloud‐based pipeline. An artificial intelligence (AI) model was trained on the publicly available Multiparameter Intelligent Monitoring in Intensive Care II (MIMIC‐II) clinical dataset, which generates a 24‐hour continuous blood pressure prediction waveform from the ECG data, and Autonomic Aging ambulatory dataset. This predicted waveform is averaged and compared with the cuff‐based blood pressure readings to assess accuracy and signal reliability. To further validate the system, cross‐device agreement in beats per minute (BPM) is used to demonstrate the plausibility of cuffless blood pressure tracking. Overall, this study presents a proof‐of‐concept framework for continuous, cuffless blood‐pressure estimation using ECG alone. By combining mechanical heartbeat verification with cloud‐based AI processing, the system demonstrates technical feasibility and provides a foundation for non‐invasive, long‐term monitoring and future clinical validation.