Deep Learning based Network based Intrusion Detection System in Industrial Internet of Things
J. Alwina Beauty Angelin, C. Priyadharsini
Abstract
The Industrial Internet of Things (IIoT) is the result of integrating the Internet of Things (IoT) into critical industries, especially in industrial and production settings. The industrial Internet of things (IoT) has many benefits, but security and privacy remain major concerns. The current intrusion detection systems (IDS) for the Internet of things have encountered a number of problems, including the inability to handle different kinds of attacks, the dependence on traditional datasets, and the lack of attention to imbalanced datasets. In addressing these challenges, this study proposes the implementation of a deep learning-based network intrusion detection system designed to identify diverse types of attacks within industrial Internet of Things environments. And to improve the effectiveness of the Network Intrusion Detection System (NIDS), the suggested model combines Convolutional Neural Network (CNN) with Auto Encoder (AE) techniques. The model improves detection outcomes by reducing data features using an autoencoder. Additionally, Convolutional Neural Networks possess the ability to automatically extract intricate features and patterns from complex data, such as network traffic. The hybrid model exhibits commendable performance on the Edge_IIoT dataset, achieving an accuracy of 92.34%, precision of 91.69%, recall of 90.28%, and an F1 score of 89.08%.