Non-cooperative Spectrum Sensing in Context of Primary User Detection: A Review

Abstract

ABSTRACTSpectrum sensing (SS) is one of the key tasks in cognitive radio networks that is performed to get awareness about the usage of electromagnetic spectrum. It enables the secondary user (SU) to detect the presence/absence of primary user (PU) and plan its transmission strategy accordingly. The aim of SS is to detect the presence or absence of PU in a certain frequency band and allow or prevent the SU's transmission on the basis of sensing result. In this paper, a review on different SS techniques is presented in the context of PU detection. We mainly focus on non-cooperative SS methods used for the PU detection. PU is the incumbent user in cognitive radio networks, and its transmission should not be interrupted. Moreover, it can start its transmission at any time, so the SU has to sense a channel before starting its transmission and monitor the arrival of PU continuously as long as it utilizes the channel to avoid any interference. The PU detection techniques are further divided into five main categories i.e. energy, covariance matrix, cyclostationary feature, matched filter, and wavelet-based detection. We investigated different algorithms based on each of these categories from PU detection perspective and discussed its various issues and challenges. Moreover, a detail comparison of the approaches is provided in terms of detection mechanism, parameters used, advantages, limitations, and comments on each scheme. Similarly, the comparison of the approaches is studied in context of detection over different levels of signal-to-noise ratio. Finally, detail findings from the study are provided for future enhancement of the PU detection mechanisms.