Intrusion Detection in a Digital Twin-Enabled Secure Industrial Internet of Things Environment for Industrial Sustainability

Abstract

Research focuses on sustainable development for the smart industry environment, where new challenges emerge every day. Digital Twins (DT) have gained substantial attention from an industrial growth point of view. This is because it significantly contributes to the predictive maintenance, simulation, and optimization of the Industrial Internet of Things (IIoT), ensuring its sustainability in future industries that demand unprecedented flexibility. Current research discusses the possibility of using DT for intrusion detection in industrial systems. By integrating DT and IIoT, physical elements become virtual representations and enhance data analytics performance. However, a lack of trust between the parties involved and untrustworthy public communication channels can lead to various types of attacks and threats to ongoing communication. With this motivation in mind, this study develops a Binary Arithmetic Optimization Algorithm with Variational Recurrent Autoencoder-based Intrusion Detection (BAOA-VRAID) for DT-enabled secure IIoT environments. The proposed BAOA-VRAEID technique focuses on the integration of DT with the IIoT server, which collects industrial transaction data and helps to enhance the IIoT environment's security and communication privacy. The BAOA-VRAID technique uses BAOA to designate an optimal subset of features to detect intrusions. The VRAE classification model with the Harris Hawks Optimization (HHO) algorithm-based hyperparameter optimizer is used for intrusion detection. The BAOA-VRAID method was tested on a benchmark dataset, showing that it significantly outperformed other contemporary methods.