Robust Spectrum Sensing Based on Phase Difference Distribution

In the era of 5G and beyond, massive wireless connectivity will be further intensified, leading to more severe spectrum shortages. Therefore, as a promising way to solve the spectrum shortage, spectrum sensing will play a more vital role. However, traditional sensing schemes are not adequate since they are sensitive to noise uncertainty and frequency mismatch. To address the problem, we propose two sensing schemes robust to noise uncertainty and frequency mismatch by leveraging the phase difference (PD) distribution. First, we derive the approximate PD distribution under Rayleigh fading channels. It consists of cosine components, which is hardly affected by noise uncertainty and frequency mismatch and differs significantly from the PD distribution of noise. Afterwards, leveraging the geometric and algebraic features of PD distribution, we propose two sensing schemes, including PD distribution shape recognition detection (SRD) and PD distribution cosine feature recognition detection (CRD), to detect spectrum holes with known CF. We further improve them for blind sensing scenarios with unknown frequency and propose the PD-based frequency estimation scheme. Simulation results show that when the signal-to-noise ratio (SNR) is lower than 0 dB, the robustness of the proposed schemes is much higher than benchmarks.