Blind Spectrum Sensing in Cognitive Radio Using Right Anderson Darling Test

Goodness of Fit tests have been used to find available spectrum with excellent detection performance in Cognitive Radio System. To extend those works, in this paper, we reformulate the spectrum sensing as a unilateral Goodness of Fit testing problem. With difference to previous available works, a random variable that obeys central F distribution with presence of primary user (PU) signal and a non-F distribution with absence of PU signal, which provides technical support for achieving blind spectrum sensing; furthermore, inspired by the thought of unilateral hypothesis test, we apply Right Anderson Darling (RAD) test to achieve bind spectrum sensing and derive a blind spectrum sensing called RAD sensing. Finally, the validness of proposed algorithm is proved by enormous Monte Carlo simulations.