Robust spectrum sensing for cognitive radio based on statistical tests

Spectrum sensing, in particular, detecting the presence of incumbent users in licensed spectrum, is one of the pivotal task for cognitive radios (CRs). In this paper, we provide solutions to the spectrum sensing problem by using statistical test theory, and thus derive novel spectrum sensing approaches. We apply the classical Kolmogorov-Smirnov (KS) test to the problem of spectrum sensing under the assumption that the noise probability distribution is known. In practice, the exact noise distribution is unknown, so a sensing method for Gaussian noise with unknown noise power is proposed. Next it is shown that the proposed sensing scheme is asymptotically robust and can be applied to non- Gaussian noise distributions. We compare the performance of sensing algorithms with the well-known Energy Detector (ED) and Anderson-Darling (AD) sensing proposed in recent literature. Our paper shows that proposed sensing methods outperform both ED and AD based sensing especially for the most important case when the received Signal to Noise Ratio (SNR) is low.