Photoplethysmography (PPG) has emerged as a highly convenient and non-invasive technique for assessing heart rate and its variability in wearable health monitoring devices. However, a drawback lies in the energy demand of PPG systems which consequently increases the average power consumption of the wearable devices integrated with PPG sensors. In this paper, we present our efforts towards packaging a novel compressed-sensing (CS) based ultra-low power PPG application-specific integrated circuit (ASIC) into a wearable device and testing it through a preclinical human study. The system comprises a custom-designed PPG analog front-end circuit, integrated with a digital back-end to implement CS, a light source-detector pair, and a commercial off-the-shelf microcontroller for Bluetooth Low Energy (BLE) based wireless data transfer. Two circular PCBs, a general-purpose main board containing the microcontroller and a plugin board housing the ASIC interfacing components, fit into a wristband form factor. This modular architecture of the wristband platform allows for the incorporation of other environmental sensors for future correlated sensing studies between health and the environment. The PPG ASIC consumes 172 $\bm {\mu }$ W power to extract heart rate from the sparse PPG signal whereas the whole system consumes 1.66 mW power for continuous streaming of heart rate data over the BLE radio which can be further duty cycled. The preclinical trial of the platform demonstrated its efficacy in assessing accurate heart rate in comparison to commercial PPG and electrocardiography systems. This work presents the first-ever demonstration of a wearable wristband with CS based PPG performed on a chip.

Wearable sleep monitoring systems provide a scalable and low-burden alternative to laboratory-based polysomnography, but overnight physiological recordings collected from wearable sensors are frequently corrupted by poor skin contact, sensor displacement, flatline behavior, saturation, abrupt autoscaling, outliers, and non-physiological noise. These low-quality segments can prevent reliable extraction of clinically relevant biomarkers, including heart rate, heart rate variability, pulse rate, oxygen saturation, and electrodermal activity, thereby limiting the robustness of downstream sleep stage classification and sleep apnea prediction. In this work, we present a multistage hybrid anomaly detection framework for signal quality assessment in a custom-designed multimodal wearable sleep monitoring platform. The proposed framework is shared across modalities: each sensor stream is segmented into windows, screened using signal processing algorithms for obvious sensor failures, and then analyzed using an autoencoder trained on normal windows to detect subtler morphology-level deviations from normal physiological patterns. On chest ECG recordings collected over 10 overnight sessions, the hybrid fusion detector achieves an accuracy of 0.9067, AUROC of 0.9068, F1 score of 0.8739, precision of 0.9470, and recall of 0.8202, outperforming rule-based detection and autoencoder-based detection alone. Additional experiments on PPG, EDA and ExG recordings show consistently high F1 scores of 0.9431, 0.9567 and 0.8728, respectively. These results demonstrate that hybrid anomaly detection can serve as an effective quality-control layer for multimodal wearable sleep monitoring and provide a foundation for future work on robust sleep stage and apnea prediction under real-world sensing conditions.

Wearable sleep monitoring systems provide a scalable and low-burden alternative to laboratory-based polysomnography, but overnight physiological recordings collected from wearable sensors are frequently corrupted by poor skin contact, sensor displacement, flatline behavior, saturation, abrupt autoscaling, outliers, and non-physiological noise. These low-quality segments can prevent reliable extraction of clinically relevant biomarkers, including heart rate, heart rate variability, pulse rate, oxygen saturation, and electrodermal activity, thereby limiting the robustness of downstream sleep stage classification and sleep apnea prediction. In this work, we present a multistage hybrid anomaly detection framework for signal quality assessment in a custom-designed multimodal wearable sleep monitoring platform. The proposed framework is shared across modalities: each sensor stream is segmented into windows, screened using signal-processing algorithms for obvious sensor failures, and then analyzed using an autoencoder trained on normal windows to detect subtler morphology-level deviations from normal physiological patterns. On chest ECG recordings collected over 10 overnight sessions, the hybrid fusion detector achieves an accuracy of 0.9204, AUROC of 0.9484, F1 score of 0.9108, precision of 0.9572, and recall of 0.8687, outperforming rule-based detection and autoencoder-based detection alone. Additional experiments on PPG, EDA, and ExG recordings show consistently high F1 scores of 0.9632, 0.9668, and 0.9779, respectively. These results demonstrate that hybrid anomaly detection can serve as an effective quality-control layer for multimodal wearable sleep monitoring and provide a foundation for future work on robust sleep stage and apnea prediction under real-world sensing conditions.