High Accuracy Respiration and Heart Rate Detection Based on Artificial Neural Network Regression

A 24GHz Doppler radar system for accurate contactless monitoring of heart and respiratory rates is demonstrated here. High accuracy predictions are achieved by employing a CNN+LSTM neural network architecture for regression analysis. Detection accuracies of 99% and 98% have been attained for heart rate and respiration rate, respectively.Clinical Relevance— This work establishes a non-contact radar system with 99% detection accuracy for a heart rate variability warning system. This system can enable convenient and fast monitoring for daily care at home.