A Weighted Linear Combining Scheme for Cooperative Spectrum Sensing

Abstract

Cooperative spectrum sensing exploits spatial diversity of secondary-users (SUs), to reliably detect the availability of a spectrum. Soft energy combining schemes have optimal detection performance at the cost of high cooperation overhead, since actual sensed data is required at the fusion center. To reduce cooperation overhead, in hard combining only local decisions are shared; however the detection performance is suboptimal due to the loss of information. In this paper, a weighted linear combining scheme is proposed in which a SU performs a local sensing test based on two threshold levels. If local test result lies between the two thresholds then the SU report neither its local decision nor sequentially estimated unknown SNR parameter values, to the fusion center. Thereby, uncertain decisions about the presence/absence of the primary-user signal are suppressed. Simulation results suggest that the detection performance of the proposed scheme is close to optimal soft combining schemes yet its overhead is similar to hard combining techniques.