Cooperative Spectrum Sensing Via Sequential Detection In Unknown-Parameter Scenarios

In this paper, we study the performance of sequential detectors (SDs) in cooperative cognitive radio networks. In the fixed sample size (FSS) detectors, the generalized likelihood ratio (GLR) test, obtained through substituting the maximum likelihood (ML) estimates of the unknown parameters in the likelihood functions, is a current alternative when deriving the uniformly most powerful (UMP) test is impossible. This idea, i.e., substituting the ML estimates of the unknown parameters in the likelihood functions, may be applicable for designing a sequential detector in the unknown-parameter problems. We consider the scenario in which some secondary users (SUs) cooperatively sense the predetermined spectrum for the detection of white spaces. In the proposed scenario, at each time instant, every secondary user sequentially estimates the unknown parameters based on all the received observations until the specific time, computes the so-called generalized log-likelihood ratio (GLLR) and sends it to the fusion center (FC). The FC combines the received statistics sequentially and determines whether to stop making measurement and creates decision. We finally express the performance of SDs in comparison to the traditional FSS detectors with the same error conditions, i.e., the probability of false alarm and missed detection.