A robust eigenvalue ratio detector for cognitive radio

Cognitive radio is a new paradigm for sharing the available radio spectrum. Using dynamic spectrum access technology cognitive radios find underutilized spectrum resources and then exploit them in an agile manner. Spectrum sensing that is used in cognitive radio systems must be reliable even in involved radio environments characterized by shadowing, fading and non Gaussian noise. This calls for robust detection algorithms that are able to operate in challenging conditions with small performance degradation. In this paper we develop a robust detector that uses the ratio of maximal and minimal eigenvalues of the input signal covariance matrix to detect the presence of primary user signal. Provided simulation results demonstrate the reliable and highly robust performance of the proposed algorithm in both Gaussian and impulsive noise environments.