Bidirectional LSTM-GRU GAN Model for Enhancing IoT Network Security

While the widespread and continuous adoption of Internet of Things (IoT) devices brings convenience to modern life, it also introduces significant security risks. Detecting malicious traffic at the packet level can significantly reduce the anomaly detection time, thereby enabling proactive security countermeasures. Using the Bidirectional Long Short-Term Memory and the Bidirectional Gated Recurrent Unit (LSTM-GRU) Generative Adversarial Network (GAN), this paper proposes a hybrid approach (denoted BiLSTM-BiGRU GAN) to enhance the capabilities of network intrusion detection systems (IDSs). The effectiveness of our approach is demonstrated by conducting a comprehensive evaluation using an IoT dataset, whose results show that incorporating GAN can significantly improve the capabilities of IDSs for detecting anomalous IoT traffic.