Exploring Edge Computing Architecture:A Hospital Workflow Efficiency Perspective

Abstract

This project presents an exploration of Edge Computing architecture, with a central focus on capitalizing on the burgeoning computational capabilities of edge devices. The motivation behind this research stems from the opportunity to harness the potential of edge intelligence architectures. This architecture leverages Edge Intelligence (EI), emphasizing machine learning inference on the edge device, and employs a cloud-fog-edge framework. Additionally, the choice of Message Queuing Telemetry Transport (MQTT) as the communication protocol is explored in detail. The research takes a hands-on approach by implementing smart pillboxes within inpatient healthcare settings. This project comprises two fundamental components. The first has a stronger software focus, involving the conception and realization of an edge computing architecture. The second has a stronger business focus, which entails quantitative analysis using simulation models to evaluate the potential time savings and efficiency enhancements using edge solutions. This research further explores the practical implications of potential time savings, particularly in the context of healthcare workflows. In doing so, it bridges the gap between edge computing theory and real-world applications from a workflow efficiency perspective. From the simulation results, the smart pillboxes yield significant time savings in medication provision. In summary, this work underscores the significance of edge computing as an enabler of efficiency using healthcare as a compelling use case to validate its transformative potential. Subject Descriptors: Software and its engineering Software organization and properties Software system structures Software architectures Hardware Communication hardware, interfaces and storage Sensor applications and deployments Applied computing Operations research Decision analysis Keywords: Software Engineering, Software and Applications, Real-Time / Embedded Systems Implementation Software and Hardware: NVIDIA Jetson Nano B01, ReactJS, ExpressJS, Flask, PostgreSQL, Picovoice