Fog-Enabled Multimodal Chest-Worn Device for Systolic Blood Pressure Monitoring

This paper presents a Fog-enabled chest-worn device for estimating systolic blood pressure (SBP). The device integrates two sensors that simultaneously detect the phonocardiogram (PCG), ballistocardiogram (BCG), and seismocardiogram (SCG) from a single location on the chest. These signals are good indicators of the early stages of cardiovascular diseases (CVD). Here, a new fog-based deep learning architecture that extracts features directly from raw signals to estimate the SBP is proposed. Three signals were transmitted to a fog computing system via a socket. Four models were implemented for analysis: Random Forest (RF), AdaBoost, Gradient Boosting (GB), and a novel deep learning architecture. The RF, AdaBoost, and GB models focused on the time delays between the main waves of the PCG, BCG, and SCG, while the deep learning architecture extracted features directly from the signals using convolutional layers. The tests involved 36 healthy volunteers under both resting and physically active conditions. The device was built using surface-mount components and powered by a 3.7 V, 250 mAh battery. The dimensions were 65 mm $\times 54$ mm, and the weight was 56 g. The three cardiac signals were simultaneously detected, with a signal-to-noise ratio higher than 41 dB. Among the four implemented models for SBP estimation in fog, the proposed deep learning architecture demonstrated the best performance, achieving an MAE of 3.5 mmHg without estimating the timing between the three detected signals. Integrating the methodology discussed here, the proposed device offers a novel method for estimating blood pressure through raw biosignals detected by a wearable device.