Investigation on identify the multiple issues in IoT devices using Convolutional Neural Network. (December 2022)

• Record Type: Journal Article
• Authors: Thouti, Swapna; Venu, Nookala; Rinku, Dhruva R.; Arora, Amit; Rajeswaran, N.
• Is Part Of: Measurement. Volume 24(2022)
• Publication Date: 2022-12
• Subjects: IoT -- CNN -- Battery -- Power supply -- Connectivity and security; Detectors -- Periodicals; Measurement -- Periodicals; 530.7
• DOI: 10.1016/j.measen.2022.100509
• Languages: English
• ISSNs: 2665-9174

Abstract

Abstract: The Internet of Things (IoT) is an innovative technology that makes it possible for physical objects like sensors, cameras, household appliances, and other objects to interact and communicate with one another. The Internet of Things (IoT) devices may exchange critical data, which makes battery/power, connectivity, and security issues crucial. An automated system for identifying and reporting abnormalities to a central controller is a prerequisite for this. In order to distinguish between approved and legitimate IoT devices, this method should be able to. IoT devices that are malicious, non-IoT devices that are malicious, and other man-in-the-middle traffic sources must all be isolated to prevent noncompliance. For improved QoS management, this aids in the formulation of administrative rules and the regulation and enforcement of network traffic. A framework-based Convolutional Neural Network (CNN) is suggested in this research to discover the aforementioned problems in IoT devices. A system that classifies IoT devices into their respective categories and reliably identifies new entries has been developed based on CNN. The findings demonstrate that CNN is capable of classifying IoT devices into the appropriate categories with the necessary accuracy as well as identifying between IoT and non-IoT devices with better accuracy.