Cooperative Spectrum Sensing for Cognitive Radio Based on Decision Tree Algorithm

Frequency spectrum is a critical resource in wireless communication systems, providing media for any communication devices from smartphones and internet-of-things devices to autonomous vehicles and smart cities. However, as the demand for wireless connectivity continues to grow, the spectrum scarcity has become a significant challenge. Cognitive radio (CR) systems have emerged as a promising solution for optimizing spectrum usage and addressing the scarcity issue. The emergence of machine learning algorithms in recent decades could offers a solution for CR systems to intelligently manage spectrum resources, allowing devices to access under-utilized spectrum. In this manuscript, we propose the decision-tree (DT) algorithm-based cooperative spectrum sensing (SS), where multiple secondary users (SUs) in CR systems cooperatively perform SS process by first estimating the power spectral density (PSD) based on the received licensed users (LUs) signals and then transmitting the PSD to a fusion center (FC). The FC integrates the PSD collected from different SUs and employs the DT algorithm to decide on the existence of LUs at different frequencies. The simulation study indicated that the accuracy of the detection on the LU presence/absence is larger than 90% except when all the ten SUs suffers from a severe fading channel with the variance of wireless fading channels tap equal to −11 dB. The LU signal and the noise power here are set to 200 mW and 50 mw, respectively.