Detection of IoT Botnet Attacks using Hybrid Deep Learning Models

Protecting interconnected IoT ecosystems requires the ability to identify botnet attacks. The proliferation of insecure Internet of Things (IoT) devices increases the risk of botnets infecting networks and facilitating the distribution of malware, theft of data, and distributed denial of service (DDoS) attacks. In order to detect attacks on the Internet of Things (IoT) botnet, this study introduces a CNN-GRU hybrid method. Utilising convolutional neural networks (CNNs) for feature extraction and gradient recurrent units (GRUs) for sequence learning, the model is able to capture spatial and temporal patterns in network traffic. The N-BaIoT dataset, which includes IoT devices infected with Mirai and BASHLITE, was used to train the model, and it has a good detection rate with little false positives. Using the Adam optimiser to dynamically change learning rates expedites detection and convergence over massive amounts of IoT data.