Resilient Detection of Cyber Attacks in Industrial Devices

Abstract

With the advent of smartphones, laptops, and home computers, smart systems are becoming more and more flexible. As the use of internet increases, there will be more cyber threats occurring on most third-party connectivity websites. The powerful technique used to detect the threats present in the IoT applications are discussed in the proposed system. Based on the KAGGLE NIDS(Network Intrusion Detection System)(Intrusion Detection System) dataset, the number of possible attacks is calculated in the proposed architecture. A similar occurrence of intrusion creating a task is detected by the system, triggering the model to prevent the intrusion by notifying the user immediately. The existing attack detection systems have a number of limitations which includes the need of human intervention to detect the attacks encountered, slower detection rate and inaccuracy in detection. An advanced deep learning algorithm is proposed for detecting possible intrusions to overcome these limitations. The proposed design focuses on creating a Novel architecture using Adaptive convolutional neural network for improving the accuracy and significantly raising the detection rate above that of the current approaches there by aiding in the immediate detection of intrusions.