DEEP LEARNING-DRIVEN TRUSTWORTHY CYBERSECURITY FOR INDUSTRIAL IOT NETWORKS

Abstract

The reliability and sustainability of the Industrial Internet of Things (IIoT) to prevent fatalities while carrying out vital tasks is a basic requirement of the stakeholders. Basic security features like trust, privacy, security, dependability, resilience, and safety are all included in a reliable IIoT-enabled network. Due to outdated security mechanism modifications, restricted update choices, and protocol variations, the conventional security processes and mechanisms are unable to safeguard these networks. Because of this, these networks need new methods to improve security and privacy measures and raise the degree of trust. In order to increase the credibility of IIoT-enabled networks, we thus suggest an innovative strategy in this study. We provide a precise and trustworthy method for detecting cyberattacks in these networks using supervisory control and data acquisition (SCADA) networks. The suggested plan integrates SCADA-based IIoT networks with deep learning-based pyramidal recurrent units (PRU) and decision trees (DT). In order to identify cyberattacks in SCADA-based IIoT networks, we also use an ensemble-learning technique. High detection rates are made possible by the ensemble DT's and PRU's nonlinear learning capabilities, which reduce the sensitivity of irrelevant features. Fifteen datasets derived from SCADA-based networks are used to assess the suggested approach. The experimental findings demonstrate that the suggested methodology works better than both conventional techniques and machine learning-based detection strategies. The suggested plan enhances IIoT-enabled networks' security and related trustworthiness metrics.