A low complexity blind multiband spectrum sensing using covariance based approach

Abstract

ABSTRACT A blind multiband spectrum sensing approach in a wideband scenario is presented using eigenvalues of a reduced sample covariance matrix formed from a small number of samples. In this approach, the wideband is split into non-overlapping multiple subbands to determine the vacant subbands. The proposed detection scheme does not require apriori knowledge about the primary users or the noise signals and has lesser computational complexity. Simulation results show better probability of detection for the proposed method in comparison with the existing methods.