Intrusion Detection in IoT Systems Based on Deep Learning Using Convolutional Neural Network

Abstract

Internet of Things (IoT) and the fourth Industrial Revolution are key developmental trends of today's technology. With a variety of devices, environments, and communication protocols, IoT systems are at increased risks of insecurity and vulnerability. Therefore, an effective intrusion detection method, which suits IoT systems, is necessary. This paper proposes a new method of detecting intrusion for IoT systems based on deep learning using a convolutional neural network. The log information of an IoT system such as location, service, address, etc., is extracted into an original feature set. Next the original feature set is improved and encoded into a digital matrix and fed into a convolutional neural network for training and detection. The proposed method is evaluated based on the cross-validation method and has an average accuracy of 98.9%.