A Cholesky decomposition and fusion clustering based spectrum sensing method

Spectrum sensing is a key technology to detect unused frequency bands, and is widely applied in spectrum sharing and dynamic channel allocation. However, it is a challenge to provide high sensing accuracy under low signal-to-noise-ratio (SNR) environments. To address this issue, this paper proposes a novel method based on feature extraction and fusion clustering. First, the sampling matrix of the received signal is decomposed into two orthogonal components I and Q, and Cholesky decomposition is performed on the covariance matrices of I and Q components to extract their two-dimensional feature vectors. Then, the fusion clustering algorithm is proposed, where the GMM clustering algorithm is performed to classify the feature vectors, and the initial parameters of GMM, such as centroids, weights and covariance matrices, are generated by K-means clustering. Simulation results show that the proposed method accelerates the convergence speed of GMM and improves the classification accuracy. It effectively enhances the performance of spectrum sensing compared to other mainstream methods.