Multiband radar signals coherent compensation with sparse representation

The coherent compensation between subband radar signal is significant for multiband fusion imaging, and it has received much attention during the past decades. At present, the most classical coherent compensation method is based on the super resolution algorithm and damped exponential (DE) model, and it works well in some experiments. However, it need the scattering center number as prescribe information which is not easy to be acquired precisely. Apart from that, the band width extrapolation (BWE) error accumulate as the band gap get wider. The cross correlation based method is computationally efficient, but its estimation accuracy depends on the sampling frequency which is decided by the hardware. In this paper, the incoherent factors between subband data are analyzed first, then due to the sparsity of the scattering centers, a sparse representation based coherent compensation method is proposed. Different from the existing methods, it does not need the scattering center number and band width extrapolation to generate frequency-band overlap data. Applying the proposed method on simulated data, the comparative experiment result indicated that the proposed method can effectively enhance the coherent compensation accuracy.