Deep Learning Based Security Preservation of IoT: An Industrial Machine Health Monitoring Scenario

Abstract

IoT creates a smart environment in which devices and sensors share information across different platforms. Most of the IoT related issues and challenges are addressed concerning applications such as smart home, smart office, smart city, smart agriculture, intelligent transportation, etc. But it is rarely addressed in terms of prognostics and health management (PHM) of industrial machinery, which is the heart of the Industry 4.0 revolution. IoT data from sensors and other ubiquitous sources, with the insights gained from Artificial Intelligence (AI), serves to transform current enterprise activities to Industry 4.0 standard. Studying the role of deep learning (DL) in IoT security preservation will be more realistic if we investigate the scope of IoT and AI in ensuring the integrity and security of devices in an industrial environment. For such a background study of IoT application and AI's intervention, this article describes the scenario of integrity and security preservation of rotating machinery, which constitutes around 40% of all the industrial machinery. An Industrial IoT(IIoT) based AI framework is proposed, which details the methods and challenges of data collection with IoT, signal acquisition-related issues in the processing phase, and different techniques and models in deep and shallow learning categories. To understand the root cause of the machinery's structural integrity issues, rotor fault diagnosis (RFD), primarily structural rotor fault (SRF), has been considered for the case study. We mainly focus on discussing the general security issues with the IIoT based on a framework developed with the RFD scenario. In short, this article investigates the role, challenges, and security issues of IoT in the industrial environment, the ways by which deep learning deals with it, etc. have been discussed in this study. © 2021, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.