Energy based Machine Learning Spectrum Sensing in 5G Cognitive Radios

A ground-breaking technology called Cognitive Radio Network (CRN) was developed to increase the effectiveness of spectrum usage. If the primary user is not using it, it gives cognitive radios the ability to transmit on the licensed portions of the spectrum. But, when the primary user decides to reclaim the spectrum, the cognitive radio must give it up. A cognitive radio contributes to more effective radio spectrum use by utilizing the untapped area of the spectrum. Spectrum sensing is the procedure used to determine the spectrum's condition. This work focuses on spectrum sensing by cognitive radios in a 5G communication network. Different ML algorithms were used for spectrum sensing and their performance was evaluated under different channel conditions such as AWGN high accuracy rates of spectrum sensing were achieved. It was observed that the logistic regression algorithm outperforms other classification algorithms at different Signal-to-Noise ratio (SNR) values for different observation window lengths. This implies that the system is able to work under different signal conditions.