Spectrum Efficient Resource Allocation Using Deep Neural Network in Underlay Cognitive Radios

In underlay cognitive radio networks, managing the primary and secondary users for resource allocation is a demanding task. This is a crucial issue that must be addressed in an efficient manner. The proposed approach, Spectrum Efficient Resource Allocation (SERA), optimizes the spectral efficiency utilizing deep neural networks in order to achieve an efficient resource allocation in underlay cognitive radios. Computation time is minimized maintaining the interference of both primary and secondary users. The results obtained by simulation show that the SERA system model offers a spectral efficiency improvement by 94.5% as compared to the IRAWCS and CASE approach.