Blind Central-Symmetry-Based Feature Detection for Spatial Spectrum Sensing

Abstract

State-of-the-art sensing methods mostly detect the spectrum holes by exploring the feature in frequency, time, and geography dimensions. In this paper, we analyze the sensing problem from angle/space domain by using the angle of arrival (AoA) estimation technology. We show a property that the spatial spectrum of noise has the feature of central symmetry, which the arriving signal does not have in general. Hence, the existence of the central symmetry feature depends on the presence or absence of the primary user signal. Motivated by this, we introduce a novel spatial spectrum sensing framework and propose a blind central-symmetry-based feature detection (CSFD) method correspondingly. Different from conventional spectrum sensing, the designed sensing framework reduces the complexity of spatial spectrum sensing and is available for spectrum access. Taking advantage of the inherent central symmetry feature of noise spatial spectrum, the proposed CSFD can achieve higher probability of detection even at low signal-to-noise ratios (SNRs) and offer AoA information. Theoretical performance analysis of the proposed CSFD method is also provided. Simulation results are presented to verify the efficiency of the proposed algorithm.