Eigenvalue-based spectrum sensing for cognitive radio

We studied the distributions of the ratio of the extreme eigenvalues and maximum eigenvalue of the complex Wishart matrix. We derived the exact decision thresholds as a function of the desired probability of false alarm for MME and MED based cooperative spectrum sensing, which are able to handle for both correlated and uncorrelated Gaussian noise samples. The expression for the decision threshold was simplified for two receiving antenna or for the case of two user collaborative sensing. We also derived a simpler closed-form threshold function using an asymptotic distribution with equal numbers of receive antennas and signal samples, that is m = n and large n. We shown that analytical and empirical results are coincide with each other. The probability of detection performance using the proposed exact decision thresholds achieve significant performance gains compared to the performance using the asymptotic decision threshold reported in the literature, which leads to efficient spectrum usage.