Detection of Temporally Correlated Primary User Signal with Multiple Antennas

In this paper, we address the problem of multiple antenna spectrum sensing in cognitive radios (CRs) when the samples of the primary user (PU) signal as well as samples of noise are assumed to be temporally correlated. We model and formulate this multiple antenna spectrum sensing problem as a hypothesis testing problem. First, we derive the optimum Neyman-Pearson (NP) detector for the scenario in which the channel gains, the PU signal and noise correlation matrices are assumed to be known. Then, we derive the sub-optimum generalized likelihood ratio test (GLRT)-based detector for the case when the channel gains and aforementioned matrices are assumed to be unknown. Approximate analytical expressions for the false-alarm probabilities of the proposed detectors are given. Simulation results show that the proposed detectors outperform some recently-purposed algorithms for multiple antenna spectrum sensing.