Skip to Main Content
An optimal spectrum sensing algorithm is proposed based on the Neyman-Pearson theorem in this paper. The probability of detection is mathematically formulated through analytical discussions. The computational complexity of proposed algorithm is lower than that of the conventional algorithm based on multitaper method with singular value decomposition (MTM-SVD) because SVD is omitted in the proposed algorithm. Secondly, we obtain the optimal number of required sensors which is adaptive to the signal to noise ratio (SNR) in every frequency bin. Thirdly, we find the proposed algorithm outperforms the conventional MTM-SVD algorithm in the detection rate, especially when the number of samples is small. The analytical assertions to performance of detection are verified by simulations. Results show that the proposed algorithm has an increase of approximately 8% and 10% in the detection rate when the number of tapers K=4 and K=6 under the constraints of SNR=0dB and the number of samples N=8, compared with the conventional MTM-SVD algorithm.