Research Article
A Cooperative Spectrum Sensing Method Based on Feature Extraction and Fusion Clustering
@INPROCEEDINGS{10.1007/978-3-031-86196-3_16, author={Jian Li and Yue Li and Xiaoxu Chen}, title={A Cooperative Spectrum Sensing Method Based on Feature Extraction and Fusion Clustering}, proceedings={Wireless and Satellite Systems. 14th EAI International Conference, WiSATS 2024, Harbin, China, August 23--25, 2024, Proceedings, Part I}, proceedings_a={WISATS}, year={2025}, month={3}, keywords={cooperative spectrum sensing Cholesky K-Means Gaussian mixture model}, doi={10.1007/978-3-031-86196-3_16} }- Jian Li
Yue Li
Xiaoxu Chen
Year: 2025
A Cooperative Spectrum Sensing Method Based on Feature Extraction and Fusion Clustering
WISATS
Springer
DOI: 10.1007/978-3-031-86196-3_16
Abstract
To improve the performance of spectrum sensing at low SNR, a collaborative spectrum sensing method based on feature extraction and fusion clustering is proposed (FEFC). First, the sampling matrix of the received signal is vectorially decomposed to obtain the I and Q component signals. Second, Cholesky decomposition is applied to the covariance matrices of the I and Q signals to fully extract their features and construct the two dimensional feature vectors. The K-Means clustering algorithm is used to optimize the initial parameters of the Gaussian Mixture Model (GMM), effectively preventing it from falling into local minima under low SNR. Finally, the feature vectors of the signals are classified using GMM clustering optimized by the K-Means algorithm to obtain the final spectrum sensing results. Simulation results show that this method reduces the convergence time of GMM and improves the accuracy of model classification. It effectively enhances the performance of spectrum sensing compared to other mainstream methods.
