Abstract

Chaotic Internet of Things (IoT) has grown quickly in the digital world. With the emergence of IoT, many security obstacles have been appearing. The study suggests an efficient intrusion detection and protection strategy utilizing hybrid artificial intelligence (AI) methods to protect IoT networks from botnet/DDoS and IP/DNS spoofing attacks. In this study, the convolutional neural network (CNN) has been proposed to detect and prevent large-scale attacks as a first shield to protect network traffic from Distributed Denial of Services (DDoS) attacks, while principal component analysis (PCA) has been proposed to detect and protect the network from malicious attacks that come from IP/DNS spoofing attacks. The results exposed 98.36% detection to protect against intrusion in doorbell IoT devices, 98.62% detection to protect against intrusion in thermostat IoT devices, and 98.81% detection to protect against intrusion in security camera IoT devices, concerning accuracy benchmarks. The CNN-PCA hybrid model was efficacious in detecting malicious and botnet attacks for numerous IoT devices with optimal security. The study compares accuracy, precision, recall, and F1-core metrics with state-of-the-art security models.