An F-test for multiple antenna spectrum sensing in cognitive radio

Spectrum sensing is an essential task in cognitive radio technology. Most of current detectors take the noise power as prior knowledge which makes the detection performance sensitive to noise uncertainty. In this paper, we propose an F-test based detector to overcome this problem. The proposed approach is robust to noise mismatch and requires low computational complexity. Based on the table of F-distribution, the exact value for the test threshold and the detection probability are derived, respectively. In addition, we shall show that the false alarm probability of the proposed method is still under control even with channel uncertainty. Simulation results demonstrate that significant performance gain achieved by the proposed approach.