An Adaptive Intrusion Detection System for IoT Networks Using Deep Learning
Pon Harshavardhanan, Venkatkumar Muneeswaran, D. Saravanan, P. Vijayakarthik, R Natchadalingam, Pundru Chandra Shaker Reddy · 2023
The rapid expansion of the Internet-of-Things (IoT) has never before attracted the attention of cybercriminals. The increasing prevalence of cyberattacks against IoT devices and intermediary communication channels lends credence to this assertion. If an IoT attack goes unnoticed for too long, it might disrupt services severely, costing money. In addition, it poses the risk of having one's identity stolen. For IoT-enabled services to be dependable, secure and financially fruitful, real-time intrusion detection on IoT devices is a must. A new intrusion detection design for IoT devices based on Deep Learning (DL) is presented in this research. In order to identify potentially harmful traffic that could lead to assaults on Internet of Things (IoT) gadgets, this smart framework employs a 4layer deep Fully-Connected (FC) network architecture. To simplify deployment, the suggested system was designed to work with any available communication protocol. During experimental performance evaluation, the suggested system has proven to be effective against both simulated and real invasions. According to the data, the proposed model is more accurate than existing methods. This novel deep learning-based IDS has highest detection rate of 95.7 percent, making it suitable for bolstering the safety of IoT networks.