Spectrum Sensing of Interleaved SC-FDMA Signals in Cognitive Radio Networks

This paper develops a spectrum sensing technique for interleaved single-carrier frequency-division multiple access (SC-FDMA) systems. By designing a metric that exploits cyclostationary features of interleaved SC-FDMA signals, we establish a framework for signal detection. Using Gaussian approximation for this metric, the parameters of the metric distributions under two hypotheses are derived, and both hypotheses are examined by the Neyman-Pearson test. We validate the accuracy of the Gaussian approximation by comparing theoretical and simulated metric histograms. The performance of the proposed method is presented for additive white Gaussian noise and multipath Rayleigh fading channels. We also investigate the effect of the block length, the number of users, the metric window length, and the presence of the pilot signals on the detection performance. Through comparative performance evaluation, we demonstrate the superiority of our proposed detection scheme over energy detection and the detection method based on autocorrelation of the cyclic prefix (CP). We obtain similar detection performance to that of the mentioned methods at about 8-13 dB lower signal-to-noise ratio (SNR). It is noteworthy that the complexity of our method is comparable to that of the energy detection and slightly higher than that of CP detection.