An IoT-Enabled Wearable Health Monitor for Synchronized ExG, PPG, and BioZ Measurement

The multimodal wearable health monitor is designed for patients with chronic conditions and leverage edge data preprocessing for efficiency. It supports a variety of use cases: a brain monitoring headband for measuring electroencephalography (EEG) and functional near-infrared spectroscopy (fNIRS), a chest health patch for sensing electrocardiogram (ECG), RR intervals (RRIs), bioimpedance (BioZ), and photoplethysmogram (PPG), and a fitness legband for recording electromyography (EMG) and PPG. Similar devices can also be connected to a body sensor network (BSN) for comprehensive health monitoring. The system includes a flexible PPG sensor and a data acquisition (DAQ) board with integrated BioZ and biopotential channels. They feature a low input-referred noise level of 25 pArms (BW =10 Hz), <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$6.06~\Omega _{\text {rms}}$ </tex-math></inline-formula> (BW =0.05–4 Hz), and <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$0.77~\mu $ </tex-math></inline-formula>Vrms (BW =0.05–150 Hz). The multimodal health monitor consumes 30.4 mW of power and can operate continuously for 36 h. The flexible sensor, with a 1-cm diameter and a bending radius of <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$66.6~\mu $ </tex-math></inline-formula>m, conforms to the skin for versatile detection. Edge data preprocessing and offloading tasks to the PC help reduce data redundancy and latency. The overall platform enables remote real-time monitoring, historical data queries, and email alerts. The results indicate that the average correlation of EEG signals was 96.21% in the eyes-open and eyes-closed tests, and the alpha rhythm was clear. During the breath-holding test, the SpO2 level changed by 10%. In chest synchronization, the heart rate (HR) increased by 2 bpm every 3 s. The mean absolute error (MAE) of the HR test for five subjects was 1.38%, and the mean absolute percentage error (MAPE) was 1.98%.