Development of a Privacy-Preserved and Secure Cooperative Spectrum Sensing System in Cognitive Radio Networks Using ATSNRNN-Enabled FPPDES With Machine Learning

Generally, Cooperative Spectrum Sensing (CSS) in Cognitive Radio Networks (CRN) recognizes the available spectrums to improve their utilization. However, CRN may lead to major security concerns. The existing studies didn’t concentrate on various privacy attacks during Spectrum Sensing (SS) in CRN. Therefore, this paper presents Format Pareto Preserving Distributed Encryption Standard (FPPDES) and Adjustable Trainable Sigmoid Nesterov-accelerated Recurrent Neural Network (ATSNRNN) - enabled privacy preserved secure cooperative SS in CRN. Initially, the primary users (PUs) and secondary users (SUs) are registered in the blockchain. Here, the Elliptic Curve Iterated Summed Cryptography (ECISC) is employed to generate the keys. Afterward, for SUs, the Fusion Center (FC) is selected using the GaussIT Skilled Optimization Algorithm (GIT-SOA) and stored in the blockchain. At this point, FC details are hashed by using Secure Lagrange Multiplied Hash Algorithm-256 (SLMHA-256). Then, spectrum sensing is performed by feature extraction, spatiotemporal evaluation, and spectrum sensing via ATSNRNN. Next, the sensing results are encrypted by using FPPDES. Afterward, the encrypted outcome is transmitted to the FC by authenticating the FC. If FC authentication verification is passed, then the encrypted information is fed to FC; otherwise, FC is reselected and updated in the blockchain. Subsequently, by using SLMHA-256, the location matching for PU and SU is done. If the location is matched, then spectrum allocation for SU is done by using GIT-SOA; otherwise, the allocation is declined. The results proved that the proposed model obtained a high accuracy of 98.88%, which outperformed the prevailing techniques.