Cuff-less Continuous Blood Pressure Estimation via Multimodal Integration of Millimeter-Wave Radar and PPG Signals

Blood pressure is a critical clinical parameter linked to various cardiac diseases. Traditional blood pressure measurement techniques often cause discomfort and are not suitable for continuous monitoring. To address these drawbacks, we utilized millimeter-wave radar for cuffless blood pressure measurement. Specifically, a 24 GHz radar signal generation module was employed to acquire photoplethysmographic (PPG) signals from a finger clip for physiological monitoring. Continuous blood pressure waveforms were recorded using the CNAP system as a reference, with synchronization achieved via the BIOPAC system. We collected physiological data from 15 subjects under three conditions: normal breathing, deep breathing, and post-exercise states, to model blood pressure. The variational mode decomposition (VMD) and finite impulse response (FIR) filter techniques were applied to reconstruct the heartbeat waveform from radar echoes. After synchronizing with the PPG signal, blood pressure was estimated using a hybrid machine learning model combining Long Short-Term Memory (LSTM) networks and Support Vector Regression (SVR). By comparing the blood pressure waveforms derived from heartbeat waveform and PPG with standard blood pressure waveforms, our results confirmed that the combination of millimeter-wave radar and finger clip achieves precise blood pressure measurements.Clinical relevance- This introduce a hybrid method of cuffless Blood Pressure measurement with acceptable accuracy.