Simple $F$ –Test-Based Spectrum Sensing Techniques for Multi-Antenna Cognitive Radios

An F-test detector with a simple analytical false alarm threshold expression is considered an alternative to the blind detectors which exhibit complicated analytical expressions. However, the existing F-test requires the channel state information (CSI) as a prior knowledge and is known to be sensitive to CSI estimation errors. In this paper, we present and evaluate simple F-test-based spectrum sensing techniques that do not require the knowledge of CSI for multi-antenna cognitive radios. Exact and asymptotic analytical performance closed-form expressions are derived for the presented detectors. Simulations assess the performance of the presented detectors and validate the derived closed-form expressions. For an additive noise exhibiting the same variance across multiple-antenna frontends, simulations also corroborate that the presented detectors are constant false alarm rate detectors which are robust against noise uncertainty.