Spectrum Sensing Based on Combined Eigenvalue and Eigenvector through Blind Learning
This work focuses on exploiting both eigenvalues and eigenvectors for spectrum sensing in cognitive radio. First, we design a blind learning algorithm for obtaining the prior knowledge of the maximum eigenvalue of noises and the leading eigenvector of primary signals by using historical sensing data. Then, we propose a new detector for spectrum sensing by exploiting both the maximum eigenvalue and the leading eigenvector. A theoretical expression for the decision threshold of the proposed detector is derived. Numerical results are provided to validate the theoretical analysis and demonstrate the superior performance of the proposed detector.