Cross-Correlation-Based, Phase-Domain Spectrum Sensing With Low-Cost Software-Defined Radio Receivers

This paper describes a highly sensitive spectrum sensing system based on a cross-correlation technique that employs low-cost software-defined radio receivers. By multiplying two data streams received by duplicated receivers and averaging the outputs, a cross-correlation system is able to suppress uncorrelated noise at the cost of measurement time. In this paper, phase-domain detection is proposed to determine whether or not a radio signal is present without any knowledge of signal features. The robustness of our method to noise uncertainty has been verified in both theoretical analyses and simulations. The sensing system has been implemented with low-cost USB tuners and a PC-based digital signal processor. TV-band spectrum sensing has been performed as a demonstration of the system.