Key points are not available for this paper at this time.
Managing the overwhelming volume of data collected by medical sensors presents a challenge in extracting relevant insights. This paper advocates for the development of an algorithm tailored to body sensor networks to identify outliers in collected data. Leveraging machine learning and statistical sampling methodologies, this research aims to optimize real-time response, particularly as computational tasks migrate to backend systems. Addressing the increasing dispersion of computing power across various domains, this study highlights the potential bottleneck posed by computation as Internet-of-Things (IoT) devices proliferate. To mitigate battery drain, a common approach involves offloading processing to background servers. However, the widespread adoption of IoT devices has sparked concerns about privacy and security. Current measures are deemed insufficient in light of escalating cyber threats. Machine learning methods offer promise in identifying vulnerabilities within IoT systems. Edge computing emerges as a solution to enhance network response times, decentralization, and security. By leveraging distributed-edge computing within an IoT framework, this paper investigates the fusion of cloud and edge computing with machine learning. Specifically, it explores how these technologies can be harnessed in the medical field, utilizing sensor-equipped IoT devices to collect extensive data for analysis. The proposed approach involves proactive decision-making at the front end, guided by an IoT server, and employs machine learning algorithms at the backend to identify pertinent data signatures. This paper underscores the significance of combining cloud, edge computing, and machine learning in a distributed-edge-computing-based IoT framework, offering a potential avenue for real-time, efficient solutions in various domains.
Sistla et al. (Thu,) studied this question.