Leveraging Global Channel State Information in Cognitive Radio IoT Networks: An IBSCNN and Red-Fox OPOSS
A. Darwin Nesakumar, K N Pavithra, T. M. Inbamalar · IETE Journal of Research · 2025
The widespread placement of sensors and devices in the Internet of Things (IoT) poses a significant challenge to spectral efficiency. Since its conception, the IoT has faced concerns about spectrum shortages. The application of Cognitive Radio (CR) technology is considered as a possible solution. CR poses challenges in the IoT, including short response times, efficient spectrum recognition in low signal-to-noise ratios. In this paper, Leveraging Global Channel State Information in Cognitive Radio IoT Networks: An Integrated Binarized Simplicial Convolutional Neural Network and Red-Fox Optimization Algorithm for Optimal Spectrum Sensing (CR-IoT-BSCN-RFOA) is proposed. Initially, data is gathered from RADIOML2016.10b dataset. Then the collected data is given into the feature extraction phase by using Refined Linear Chirplet Transform (RLCT) for extracting features, like spectrum holes detection, signal entropy, power spectral density, SNR, adjacent user range, received signal strength index, bandwidth, fading fact. The Binarized Simplicial Convolutional Neural Network (BSCN) is used to predict optimal primary user. The Red-Fox Optimization Algorithm (RFOA) is proposed to optimize the weight parameters of BSCN. Finally, the performance of the proposed CR-IoT-BSCN-RFOA method provides 27.10%, 29.22% and 28.36% higher accuracy when compared with existing techniques: Fuzzy ELM-dependent optimal spectrum sensing in CR-Internet of Things network (OSS-CR-IoT), Multi-Antenna Spectrum Sensing along Alpha-Stable Noise for Cognitive Radio-aided IoT (MAS-CR-IoT) and Adversarial Attacking and Defensing Modulation Recognition with DL in Cognitive Radio-assisted IoT (ADMR-DL-IoT) methods respectively.