A Spectrum Sensing Algorithm for Constant Modulus Primary Users Signals

In this letter, a spectrum sensing technique is proposed for constant modulus primary users signals. A new metric, namely the normalized variance, is introduced as a function of the second and fourth order moment of the received signal. Then, it is used as the testing variable for detecting the primary user signal. Theoretical analysis and simulations in the presence of noise uncertainty evaluate the performance of the proposed technique, in comparison with the conventional energy detection. The results evidence the efficiency of our method for fast and reliable primary users’ detection in cognitive radio systems.