Optimized spectrum sensing and security enhancement for cognitive radio internet of things

Abstract

The rapid growth in wireless communication services has led to an increasing demand for high bandwidth and data rates, raising concerns about potential spectrum exhaustion in the near future. This research addresses critical challenges in Cognitive Radio Networks, particularly the limitations of fixed thresholds in existing spectrum sensing methods and the vulnerability of collaborative sensing to Byzantine attacks. We propose three innovative techniques aimed at enhancing spectrum sensing, ensuring security, and improving spectral efficiency. First, we introduce an Energy Detection with Non-Parametric Amplitude Quantization method, optimized using an Arithmetic Optimization Algorithm. This approach effectively establishes optimal thresholds based on Constant False Alarm Rate and Constant Detection Rate principles, significantly improving detection reliability and providing better protection for Primary Users. Second, we tackle the security challenges in Cognitive Radio Networks by developing a solution that employs an optimized Dual-Channel Capsule Generative Adversarial Network and Auto-Metric Graph Neural Network. This method is designed to detect and mitigate Byzantine attacks, thereby preserving the integrity of the collaborative spectrum sensing process and enhancing overall network security. Lastly, we address spectral efficiency issues in Internet of Things applications through the Global Channel State Information approach, newline