Trust Determination-based Cybersecurity in Cognitive Radio Networks using Transformer-Enabled Residual Bidirectional RNN
P. Shyamala Bharathi, A Shankar · 2025
The rapid development of new services and technologies in the wireless domain increases the spectrum scarcity, cognitive radio solves this issue by sharing the licensed bands without providing any harmful interference to licensed users. Cognitive radio technology provides promising solutions in terms of robust cross-layer adaptation, dynamic spectrum sharing and collaborative networking. Secure communication in cognitive networks is crucial because the dynamic behaviour and the vulnerabilities present in the network produce corruption, risk of losses, theft of data, and inappropriate changes. Trustworthiness of the secondary user is important to maintain integrity, confidentiality, and availability of information and systems. To solve the issue of unauthorized usage of spectrum bands, and selfish misbehaviours, a deep learning-based effective security management model is implemented in this work based on the trust determination between the neighbouring nodes. The cyberattacks are detected using the proposed deep learning model for providing sufficient security in the cognitive radio networks. Here, the required data are initially collected from the available data resource. The gathered data are directly subjected to the developed model called Transformer Residual Bidirectional Recurrent Neural Network (TResBi-RNN). This network distinguishes the malicious nodes to assure confidentiality. The experimental analysis was conducted on the developed model compared with other classical techniques. From the result, it is concluded that the initiated model achieved higher and advanced cybersecurity in cognitive radio networks.