Area-Based Eigenvalue Spectrum Sensing for CRNs Using Marchenko–Pastur Bounds

A novel eigenvalue-based spectrum sensing technique for cognitive radio networks is presented. The method leverages the behavior of a Wishart matrix characterized by Marchenko-Pastur bounds based on Random Matrix Theory (RMT) to construct a robust detection framework. Unlike conventional methods that rely solely on extreme eigenvalues, the proposed approach utilizes the entire empirical eigenvalue distribution within the Marchenko–Pastur bounds. By integrating the probability density function of eigenvalues estimated via kernel density estimation, we construct a test statistic that significantly enhances detection performance, particularly under low Signal-to-Noise Ratio (SNR) conditions. Simulation results demonstrate superior performance compared to traditional energy detection, with improved robustness to noise uncertainty and flexibility in balancing detection accuracy and computational efficiency.