A cooperative spectrum sensing method based on signal decomposition and K-medoids algorithm

To solve the problem of low sensing performance and low accuracy of threshold estimation in traditional spectrum sensing systems with low signal-to-noise ratio (SNR), we proposes a cooperative spectrum sensing (CSS) method based on signal decomposition and K-medoids clustering algorithm. Firstly, to improve the sensing performance of the system in the case of fewer cooperative secondary users, a feature extraction method based on empirical mode decomposition and matrix decomposition and recombination is proposed. The method can accurately acquire the characteristic information of the sampled signal and improve the feature accuracy. Finally, the features are classified using the K-medoids clustering algorithm. In the experimental part, the result shows that the method can effectively improve the sensing performance of the spectrum sensing system at low SNR.