Antenna selection based spectrum sensing for cognitive radio networks

In a cognitive radio system, the accuracy of spectrum sensing is critical for both primary and secondary systems. Both cooperative sensing and single-user multiple-antenna techniques have been proposed in the literature to enhance the accuracy of spectrum sensing. However, these techniques may require additional signaling and hardware costs, and can lead to user reliability problems or energy consumption issues. To avoid these problems, we propose a scheme that requires no extra signaling and low hardware costs, while still offering substantial diversity gains for enhancing the sensing performance. This is achieved by combining antenna selection with the spectrum sensing mechanism. We analyze the diversity order of the proposed scheme and show that no loss occurs relative to the case of single-user multiple-antenna sensing. This benefit is achieved while keeping a low implementation complexity as only a subset of RF chains are used in a given time period. Simulations show that a substantial sensing gain can be obtained compared to using other traditional multiple-antenna techniques in some practical cases.