A Selective-Relay Based Cooperative Spectrum Sensing Scheme without Dedicated Reporting Channels in Cognitive Radio Networks

Abstract

Typically, each cooperative spectrum sensing process requires two phases: the primary user's signal detection phase, in which all cognitive users attempt to detect the presence of the primary user within a certain observation window (called signal detection overhead); and the initial detection result reporting phase, in which the cognitive users forward their detection results to a fusion center. To avoid interfering with the primary user in the reporting phase, previous research assumed that there is a common control channel (also known as dedicated reporting channel) between the cognitive users and fusion center, which, however, requires extra channel resources and introduces an additional complexity due to the dedicated channel resource management. In this paper, we propose a selective-relay based cooperative spectrum sensing scheme, which is able to control and reduce the interference from cognitive reporting users to primary user without the dedicated channel. We analyze the interference impact on the primary user and show that the interference induced by the reporting users is controllable and can be reduced to satisfy a given outage probability requirement of the primary transmissions. In addition, we investigate the receiver operating characteristics (ROC) of the traditional cooperative sensing scheme (with dedicated reporting channel) and the proposed scheme (without dedicated reporting channel) by jointly considering the signal detection and reporting phases. It is proven that, given a target detection probability, a unique optimal signal detection overhead exists to minimize an asymptotic overall false alarm probability in high SNR regions. We illustrate that, compared to the traditional scheme, the selective-relay based cooperative sensing scheme can save the dedicated channel resources without sacrificing ROC performance. Numerical results also show that, under a guaranteed overall detection probability, an overall false alarm probability can be minimized through an optimization of the signal detection overhead.