Noncoherent Diversity Combining for Spectrum Sensing

Spectrum sensing is a key enabling function for opportunistic spectrum access in a cognitive radio system. In this work, the spatial diversity of multiple spatially distributed antennas is exploited, and diversity combining schemes are studied to improve the spectrum sensing performance. In contrast to existing diversity combining techniques, the proposed detectors do not rely on knowledge of realizations of channel fading and hence operate in a noncoherent mode. The optimal detector developed combines the received signals from multiple antennas based on statistical knowledge of channel fading. Two simplified detectors are also proposed that trade off between implementation complexity and sensing performance. For all the presented detectors, closed-form expressions of the probability of false alarm and the probability of miss detection are derived, and are investigated through numerical examples under the Neyman-Pearson criterion.