Broadband Spectrum Sensing Based on Exponential Sum Decomposition

In this paper, a blind broadband spectrum sensing based on Hankel matrix and Kung method is proposed. Specifically, first, the entire broadband spectrum is sampled by a sampling filter, and sampled signals are transformed into a Hankel matrix to display signal subspace. Second, using singular value decomposition (SVD) technique to construct the Hankel matrix, and the number of occupied subbands can be evaluated. Third, centers of the occupied subbands are estimated by applying the Kung method. The proposed method can determine the number and the centers of occupied subbands at the same time. Simulations show that the proposed method has a great detecting performance, especially in low signal-to-noise ratio (SNR) region. Moreover, complexity of the method is also relatively low.