부산대학교 도서관

Some bands in the frequency spectrum became overloaded and others underutilized due to considerable increase in demand and user allocation policy. Cognitive radio applies sensing techniques to dynamically allocate secondary users. Cooperative spectrum sensing are currently showing promising results. So, in this work, we propose a cooperative spectrum sensing system based on a residual convolutional neural network architecture. For the proposed method, accuracy was more than 95% for noise power density below -124dBm/Hz, with the cooperation of 10 secondary users.