A Review of Consumer-Grade Wrist-Worn Wearables for AI-Based Health Monitoring and Early Warning

Abstract

Consumer-grade wrist-worn wearables have recently arisen as a promising tool for daily life health detection; however, their accuracy and bias remain under debate. In this review, the author discussed the latest development of consumer-grade wrist-worn wearables in detecting OSA (obstructive sleep apnea) and AF (atrial fibrillation), by using PPG (Photoplethysmography), SpO₂ (oxyhemoglobin saturation), and IMU (inertial measurement units). Based on these two illnesses, the author evaluates how skin color may influence the OSA and AF detection. The author pointed out the findings for OSA and AF detection in different sensors and approaches. For example, in OSA, the author analyzes the method of detecting changes in SpO₂ and changes in posture by using the IMU. In AF detection, the author analyzes the two-step approach of AF capturing, analyzes the statistical data for On-demand ECG confirmation, and studies experiments from Apple Inc. and Huawei Inc. Based on that, the author shows three future perspectives, such as improvements that could be made in the device and users, and the supervision department, then summarizes the entire text.