Deep Learning based Spectrum Utilization in Cognitive Radio Networks using ASAR Algorithm

Abstract

In cognitive radio networks (CRN), primary users (PU) are the licensed users and secondary users (SU) are the unlicensed users. PUs can use the spectrum in the primary network (PN) which is licensed, but the SU can only use the spectrum that is not used by the PU. Also, the SU should not affect the transmission of PU. Thus, the spectrum allocation for SU is a major problem in CRN. Therefore, in this paper, cooperative spectrum sensing (CSS) is considered with adaptive search and rescue (ASAR) optimization and auto-encoder deep learning. For the effective sharing of spectrum, a novel classification method based on deep learning is proposed. Here, for the detection of PU, multiple SUs are considered. The implementation tool used for the proposed method is NS2. Assessment metrics are performed based on delay, throughput, network lifetime, etc. Also, the balance between the transmitted power and interference is maintained to achieve high throughput.