Advanced Radar-Based Non-Contact Blood Pressure Monitoring: A Machine Learning Approach

The pursuit of non-invasive healthcare monitoring has led to the development of bioradar technologies for blood pressure assessment. This study introduces an innovative hybrid computational approach that integrates Convolutional Neural Networks (CNN), Bidirectional Long Short-Term Memory networks (BiLSTM), and an attention mechanism for precise, continuous measurement of systolic and diastolic blood pressure. Using a physiological dataset collected from 30 subjects with a 24GHz radar, we validated the performance of the proposed method. The accuracy of our method surpassed that of standalone CNN and CNN-LSTM hybrid models. Moreover, the Mean Estimation Error (ME) and Standard Deviation (STD) of our method on the static resting scenario dataset comply with the stringent standards set by the Association for the Advancement of Medical Instrumentation (AAMI). The results indicate that the proposed method significantly enhances the model's predictive capabilities, offering a promising solution for non-contact blood pressure monitoring.