Wearable Health Monitoring: Challenges and Opportunities

With the advent of wearable biomedical sensors and edge computing using low power microcontrollers, ambulatory health monitoring systems have been revolutionized over the last decade. Typical applications include cardiac arrhythmia detection, heart rate measurement and respiration tracking. The system consists of miniature biomedical sensors connected to a low power microcontroller which continuously analyzes the signals for prediction of abnormal events and also streams the data to a handheld device or IoT-enabled gateway for data archiving. The primary design considerations of these systems are energy consumption, form factor, complexity of abnormality detection and communication. Electrocardiogram (ECG), galvanic skin response (GSR) and photoplethysmogram (PPG) have been the principal sensing modalities of these systems. The embedded computing flow consists of signal quality assessment, filtering, features extraction and prediction (either abnormality, or a new parameter). Since resource constrained hardware is used, there should be an optimization between the desired detection accuracy and complexity of the detection algorithm. Data compression over the communication link can further enhance the energy efficiency of the overall system. This article provides a brief snapshot of wearable health monitoring systems with special emphasis on one-device computing.