Predicting Energy Consumption Through Machine Learning Using a Smart-Metering Architecture

Abstract

Extensive Internet of Things (IoT) networks consisting of billions of smart interconnected devices can serve a plethora of functions. The scale of these networks poses several architectural challenges, especially when combined with the essential requirements of reliable device telemetry, automated remote management, and multilayer security. In this article, we outline a flexible smart-metering architecture that can provide device monitoring and management in a unified manner over disparate underlying network technologies, such as narrow-band IoT (NB-IoT), LTE-Cat- M1, Zigbee, Wi-Fi, Wireless Smart Ubiquitous Network (Wi-SUN), longrange wide area network (LoRaWAN), and Sigfox.