Electro-Mechanical Data Fusion for Heart Health Monitoring

Abstract

Heart disease is a major public health problem and one of the leading causes of death worldwide. Therefore, cardiac monitoring is of great importance for early detection and prevention of adverse conditions. Recently, there has been extensive research interest in long-term, continuous, and non-invasive cardiac monitoring using wearable technology. Here we introduce a wearable device for monitoring heart health. This prototype consists of three sensors to monitor electrocardiogram (ECG), phonocardiogram (PCG), and seismocardiogram (SCG) signals, and a microcontroller module with Bluetooth wireless connectivity. Our preliminary results show that the device can record all three signals in real time. In our initial attempt at signal processing, a recurrent neural network (RNN) based machine learning algorithm, Long Short-Term Memory (LSTM), is used to monitor and identify key features in the ECG data. The next phase of our research will include cross-examination of all three sensor signals, development of machine learning algorithms on PCG and SCG signals, and continuous improvement of the wearable device.