An Evolution Framework for Evaluating Intelligent Spectrum Sensing Mechanisms in Cognitive Radio

Cognitive radio technology has advanced to address the challenge of limited spectrum availability by employing spectrum sensing techniques. This process is framed as a detection problem involving two core hypotheses: Hypo-0 and Hypo-1. Hypo-0 represents the scenario where only noise is present, and its modeling depends on parameters such as the number of observed samples, the timing of signal acquisition, and the presence of Gaussian white noise with a defined variance. In contrast, Hypo-1 accounts for the presence of a signal, incorporating a channel propagation factor that reflects the interference power received. Several techniques have been developed to address this challenge, with energy detection emerging as a widely adopted approach due to its ability to function without requiring prior knowledge of the signal. However, its effectiveness diminishes under low signal-to-noise ratio (SNR) conditions. This paper introduces an evaluation framework that formulates the spectrum sensing problem using two primary hypotheses, H0 and H1. It benchmarks conventional energy detection techniques and those specifically designed for low SNR environments through comprehensive statistical analysis. This work introduces a novel framework for benchmarking spectrum sensing methods, filling a gap not addressed in current literature. It evaluates the performance of ten cooperative spectrum sensing techniques, Maximum Eigenvalue Detection (MED), Generalized Likelihood Ratio (GLR), Maximum Minimum Eigenvalue Detection (MMED), Energy Detection (ED), Arithmetic to Geometric Mean Ratio (AGM), Hadamard Ratio (HR), Volume-Based Detection (VD), Gershgorin Radii Centers Ratio (GRCR), Gini Index Detection (GID), and the newly proposed Rician Rice Factor Based Detection (RFD). The analysis focuses on the probability of detection across varying signal-to-noise ratio (SNR) levels, from low to high, as well as different numbers of secondary or cognitive users. Among all methods evaluated, the proposed RFD technique consistently achieves higher detection accuracy, maintaining strong performance under both challenging noise conditions and with an increasing number of cognitive users.