Enhancing Security in Next Generation Networks: A Deep Learning Approach for Intrusion Detection
V Saraswathi, R. Dimple Dayana · 2025
The evolving landscape of 6G networks demands advanced solutions to address critical security and privacy challenges, particularly against the increasing sophistication of cyber-attacks. Traditional Intrusion Detection Systems (IDS) struggle to address the evolving and intricate nature of these cyber threats, often resulting in high false positive and false negative rates. Recent techniques, such as access control mechanisms, firewalls, and encryption, have limitations in providing comprehensive protection in 6G environments, especially in cases of denial-of-service attacks. This research focuses on developing a robust IDS tailored for 6G networks, leveraging the Long Short-Term Memory Recurrent Neural Network (LSTM-RNN) model to detect evolving attack patterns. To mitigate gradient vanishing issues, the NADAM optimizer is integrated, enabling superior performance with an efficiency score of 9.45% and a false alarm rate (FAR) of 9.85%. The proposed framework outperforms traditional optimizers such as RMSprop, Adagrad, and Adam, offering improved reliability and precision in detecting intrusions. Additionally, future perspectives on integrating blockchain for secure spectrum and data sharing in 6G networks are discussed. This research emphasizes the potential of deep learning and blockchain to develop secure, efficient, and resilient next-generation communication systems.