Weighted Correlation-Based Spectrum Sensing for Cognitive Radio in Rayleigh Fading Channels

Abstract

Correlation-based algorithms are low-complexity spectrum sensing methods requiring little knowledge on primary signals or noise signals. However, their detection performance severely degrades in the low signal-to-noise ratio (SNR) regime with low signal correlation, which happens to be quite common in practice. In this paper, a weighted correlation- based spectrum sensing scheme and its simplified blind detection scheme are proposed to effectively improve the detection performance. The two proposed algorithms adequately exploit both the auto- correlation and the cross-correlation function statistical characteristics of received signals and assign a proper weight to each term in the test statistics of them, enlarging the differences of test statistics with or without the existence of primary users (PUs), making it easier to detect PUs and thus greatly promoting the detection performance. The weighting operation is verified to be effective from two aspects. The two proposed algorithms are proved robust against noise uncertainty. The false alarm probabilities and detection probabilities are analyzed in the low SNR regime, and their approximate analytical expressions are derived based on the central limit theorem. The analyses are verified through simulations. Experiments with simulated multi-antenna signals show that the proposed detectors can significantly outperform other correlation-based detectors with about 2.2-dB detection performance gain.