A LOW-COMPLEXITY SUB-NYQUIST BLIND SIGNAL DETECTION ALGORITHM FOR COGNITIVE RADIO

The detection of sparse wideband signal in the sub-Nyquist regime is considered in this paper. We present a low-complexity and robust multiband signal detection algorithm based on algebraic analysis and statistical methods. The original signal is subsampled with Multi-coset sampling. We find that there are some linear constraints between the nonzero spectrum locations. The linear relationship is described by a frequency locator polynomial. The detector does not require priori knowledge about the frequency locations of the signals of interest. Moreover, we show that our method has lower complexity of both samples and computation compared with cyclostationary detection (CD) in the sparse case. Numerical results demonstrate our detector outperforms energy detection (ED) in the sub-Nyquist regime especially in low signal to noise ratio (SNR).