Lightweight Hybrid Deep Learning Architecture and Model for Security in IIOT

Abstract

Remarkable progress in the Internet of Things (IoT) and the requirements in the Industrial era have raised new constraints of industrial data where huge data are gathered by heterogeneous devices. Recently, Industry 4.0 has attracted attention in various fields of industries such as medicines, automobiles, logistics, etc. However, every field is suffering from some threats and vulnerabilities. In this paper, a new model is proposed for detecting different types of attacks and it is analyzed with a deep learning technique, i.e., classifier-Convolution Neural Network and Long Short-Term Memory. The UNSW NB 15 dataset is used for the classification of various attacks in the field of Industry 4.0 for providing security and protection to the different types of sensors used for heterogeneous data. The proposed model achieves the results using Cortex processors, a 1.2 GHz processor, and four gigabytes of RAM. The attack detection model is written in Python 3.8.8 and Keras. Keras constructs the model using layers of Convolutional, Max Pooling, and Dense Layers. The model is trained using 250 batch size, 60 epochs, 10 classes. For this model, the activation functions are Relu and softmax pooling.

Development of lightweight intrusion model in Industrial Internet of Things using deep learning technique

“…Furthermore, a DL procedure, namely a Classifier-Convolution Neural Network Memory [24], and a rule-based featureselection algorithm [25] have also been used to provide a unique model for recognizing different kinds of attacks. Applying deep neural networks (DNN) with bidirectional LSTM as a hybrid approach in real-time poses challenges in terms of accuracy [24].…”

“…Furthermore, a DL procedure, namely a Classifier-Convolution Neural Network Memory [24], and a rule-based featureselection algorithm [25] have also been used to provide a unique model for recognizing different kinds of attacks. Applying deep neural networks (DNN) with bidirectional LSTM as a hybrid approach in real-time poses challenges in terms of accuracy [24]. This is due to the lack of validation on dynamic datasets and the requirement for more reliable results when using various kinds or mixes of deep learning techniques on diverse datasets.…”

“…The existing techniques, such as KNN [23], CNN-LSTM [24], Bi-LSTM [28], HDRaNN [26], and proposed technique are tested with the dataset, namely DS2OS, to reveal the effectiveness of the proposed model. The results, as obtained with existing and proposed techniques, are shown in Figs.…”