Distributed Analytics and Edge Intelligence

Abstract

Biomedical research and clinical practice are entering a data-driven era. One of the major applications of biomedical big data research is to utilize inexpensive and unobtrusive mobile biomedical sensors and cloud computing for pervasive health monitoring. However, real-world user experiences with mobile cloud-based health monitoring were poor, due to the factors such as excessive networking latency and longer response time. On the other hand, fog computing, a newly proposed computing paradigm, utilizes a collaborative multitude of end-user clients or near-user edge devices to conduct a substantial amount of computing, storage, communication, and etc. This new computing paradigm, if successfully applied for pervasive health monitoring, has great potential to accelerate the discovery of early predictors and novel biomarkers to support smart care decision making in a connected health scenarios. In this paper, we employ a real-world pervasive health monitoring application (pervasive fall detection for stroke mitigation) to demonstrate the effectiveness and efficacy of fog computing paradigm in health monitoring. Fall is a major source of morbidity and mortality among stroke patients. Hence, detecting falls automatically and in a timely manner becomes crucial for stroke mitigation in daily life. In this paper, we set to (1) investigate and develop new fall detection algorithms and (2) design and employ a real-time fall detection system employing fog computing paradigm (e.g., distributed analytics and edge intelligence), which split the detection task between the edge devices (e.g., smartphones attached to the user) and the server (e.g., servers in the cloud). Experimental results show that distributed analytics and edge intelligence, supported by fog computing paradigm, are very promising solutions for pervasive health monitoring.