Main Article Content
The following four primary steps (a) pre-processing, (b) feature extraction, (c) attack detection, and (d) attack mitigation) are used to create a unique IIOT attack detection and mitigation framework in this research work. The acquired raw data (input) is first treated to a pre-processing step, which includes data normalization activities. subsequently, the features retrieved from the pre-processed data include technical indicators, enhanced higher order statistical features (Skewness, Kurtosis, Variance, and Moments), and improved Mutual Information, Symmetric Uncertainty, Information gain ratio, and Relief based features. A two-stage ensemble of classifiers is used to build the attack detection framework, which comprises the “Gated Recurrent Unit (GRU), Recurrent Neural Network (RNN), Convolutional Neural Network (CNN), and Optimized Deep Belief Network (DBN)”. The retrieved features are used to train the Gated Recurrent Unit (GRU), Recurrent Neural Network (RNN), and Convolutional Neural Network (CNN) that resides within the first stage of the ensemble-classifier. The optimal Deep belief Network (DBN)-in the second layer of ensemble classifier, which is trained with the outcomes obtained from the “Gated recurrent unit (GRU), Recurrent Neural Network (RNN)”, and Convolutional neural Network, determines the final detection about the presence/absence of attack in the IIoT network (CNN). The weight functions of the Deep belief Network (DBN) are optimized utilizing the newly projected Migration updated with Supervisor guidance (MUSG) to obtain greater detection accuracy. The proposed hybrid optimization model combines both the Sandpiper Optimization Algorithm (SOA) and the Teamwork Optimization Algorithm (TOA) concepts. The control is handed to the attack mitigation framework whenever an attacker is discovered within the network by the optimised Deep belief Network (DBN). Attack Mitigation Framework: Using the updated BAIT technique, the discovered attacker is mitigated. As a result, the IIoT network is protected to be efficient via comparative analysis.