Cooperative Spectrum Sensing in Cognitive Radio Networks via an Adaptive Gaussian Mixture Model Based on Machine Learning

Abstract

Spectrum resources are becoming extremely scarce in modern wireless communication. However, the majority of the currently available spectrum resources have not been fully utilized. To mitigate this problem, we suggested Machine learning-based Adaptive Gaussian Mixture Model (AGMM) for cooperative spectrum sensing in cognitive radio networks for pattern classification. We employ the energy level of secondary users to build a feature vector in the proposed method. The training feature vectors for classification are well defined by a combination of Gaussian density functions that are obtained using the proposed method. The proposed method performance is evaluated in terms of accuracy, recall, F1 score, and Receiver Operating Characteristics (ROC) curves. The performance parameters of the proposed method are compared to the existing K-mean clustering approach. As evidenced by the results, the proposed method performs better than an existing method in all comparison parameters, according to the simulation findings in the MATLAB version.