Imperfect full duplex spectrum sensing in cognitive radio networks

Abstract

Time-slotted, which means that primary users only change their status (active or not active) at the start of each secondary frame, has been considered as a common assumption in cognitive radio networks (CRNs). However, in realistic cases, primary users are non-time-slotted, which means primary users can be active or not active at any time during the whole secondary frame duration. In not-time-slotted CRNs, it is difficult to obtain good performance by using the traditional half duplex spectrum sensing scheme. In this paper, we propose a novel full duplex spectrum sensing scheme for non-time-slotted CRNs. We derive the probability of detection and the probability of false alarm with random arrival/departure of primary users' traffic and develop a continuous time Markov chain model for non-time-slotted CRNs. We analyze the effect of bandwidth, antennas placement error, and transmit signal amplitude difference on the performance of non-time-slotted CRNs. Numerical results show that the bandwidth, which is the most unavoidable imperfect factor, has little impact on performance of non-time-slotted CRNs. Therefore, the full duplex spectrum sensing scheme can be effectively used in wideband CRNs.