A Survey on Cognitive Radio Network Spectrum Sensing using Deep Learning Methods

A potential novel strategy for addressing the increasing demand for spectrum sensing in cognitive radio networks (CRNs) involves utilizing underutilized frequency bands through spectrum sensing. Spectrum sensing, a fundamental component of CRNs, detects available spectrum bands without disrupting signals from licensed users. Spectral sensing has recently emerged as a domain for deep learning algorithms, which have the capability to autonomously differentiate complex patterns and representations from raw spectrum data. This study conducts a comprehensive review of the state-of-the-art in deep learning for CRN spectrum sensing. It begins with an introductory overview of spectral sensing in CRNs and proceeds to assess several attempts focusing on the integration of deep learning into spectrum sensing. Further, the paper explores into crucial research challenges and potential opportunities associated while employing deep learning for spectrum sensing in CRNs. These challenges encompass issues such as data scarcity, adaptability to diverse environments, energy efficiency, resilience to interference and noise, and real-time processing constraints. The utilization of deep learning for spectrum sensing in CRNs represents an emerging field, and this review includes current methodologies, challenges, and potential remedies in this domain.