Statistical spectral analysis: a nonprobabilistic theory
Statistical spectral analysis: a nonprobabilistic theory
Hi-index | 0.00 |
Detecting the presence of primary users in a licensed spectrum is the very task upon which the entire operation of cognitive radio rests. The paper proposes a fast novel spectrum sensing algorithm for cognitive radios based on cyclic autocorrelation. The method is founded on the basic mathematical model of digitally modulated signals which can be ASK, PSK, QAM etc. When only the existence of primary users in noise is detected, special cyclic frequency α = 0 can be choosed to sense, which will significantly reduce the computational cost in applying the cyclostationarity detection. In this way it is easily applicable because it is also a blind detection method. It outperforms the energy detector in the presence of noise power uncertainty. Derivation and analysis for the proposed algorithm is given. Computer simulations are presented to verify the method.