A spectrum sensing algorithm based on correlation coefficient and K-means

In order to improve the detection probability in the environment of low signal-to-noise ratio (SNR), and solving the problem of complex threshold derivation in traditional spectrum sensing technology, the improved spectrum sensing method is proposed in this paper. Firstly, the signal received by each secondary user is decomposed and recombined (DAR). Then the correlation coefficient (CC) based on the sampling signal matrix is extracted as the decision statistic, which reduces the influence of the noise uncertainty. Finally, the K-means clustering algorithm is used to class these decision statistics, accuracy greatly. In order to facilitate expression, the proposed algorithm is abbreviated as DARCCK. Through experimental simulation, the DARCCK algorithm exhibits better detection performance than the energy detection (ED), the maximum and minimum eigenvalue (MME) algorithm and the difference between the maximum and minimum eigenvalues (DMM) in the communication environment with low SNR.