Digital orthogonal receiver is one of the key techniques in digital receiver of soft radar, and compressed sensing is
attracting more and more attention in radar signal processing. In this paper, we propose a CS digital orthogonal
receiver for wideband radar which utilizes compressed sampling in the acquisition of radar raw data. In order
to reconstruct complex signal from sub-sampled raw data, a novel sparse dictionary is proposed to represent the
real-valued radar raw signal sparsely. Using our dictionary and CS algorithm, we can reconstruct the complex-valued
radar signal from sub-sampled echoes. Compared with conventional digital orthogonal radar receiver, the
architecture of receiver in this paper is more simplified and the sampling frequency of ADC is reduced sharply.
At the same time, the range profile can be obtained during the reconstruction, so the matched filtering can
be eliminated in the receiver. Some experiments on ISAR imaging based on simulated data prove that the
phase information of radar echoes is well reserved in our orthogonal receiver and the whole design is effective for
Radar imaging is an ill-posed linear inverse problem and compressed sensing (CS) has been proved to have
tremendous potential in this field. This paper surveys the theory of radar imaging and a conclusion is drawn
that the processing of ISAR imaging can be denoted mathematically as a problem of 2D sparse decomposition.
Based on CS, we propose a novel measuring strategy for ISAR imaging radar and utilize random sub-sampling
in both range and azimuth dimensions, which will reduce the amount of sampling data tremendously. In order
to handle 2D reconstructing problem, the ordinary solution is converting the 2D problem into 1D by Kronecker
product, which will increase the size of dictionary and computational cost sharply. In this paper, we introduce the
2D-SL0 algorithm into the reconstruction of imaging. It is proved that 2D-SL0 can achieve equivalent result as
other 1D reconstructing methods, but the computational complexity and memory usage is reduced significantly.
Moreover, we will state the results of simulating experiments and prove the effectiveness and feasibility of our