Through-wall Radar Imaging Research Based On Group Sparse Compressed Sensing

Through-the-Wall Radar has the ability to detect and image targets hidden behind a wall or inside a building.It has a wide range of applications in post-disaster rescue,counter-terrorism operations and many other scenes.As a new radar imaging technology applied in complex environments,Through-the-Wall Radar Imaging(TWRI)still face challenges in many respects.On the one hand,high-resolution imaging requirements increase system complexity(such as large bandwidth,multiple antennas,etc.)which leads to a large number of measurement data.It imposes a great burden on the system's storage,transmission,processing procedures.On the other hand,due to the reflection of electromagnetic waves by the wall,the receiver receives target returns from multiple propagation paths.It generates so-called multipath effect that results in unexpected "ghosts"in the reconstructed image and affects the quality of image reconstruction.To deal with these problems,this thesis mainly studies the high-resolution imaging and multipath suppressing technologies based on the compressive sensing(CS)theories.The studies are based on group sparsity observed in the received signals of through-the-wall radar.In order to exploit the sparsity of through-the-wall radar,Bayesian compressive sensing(BCS)is utilized to achieve high-resolution imaging with low data volume.The multipath effect of through-the-wall radar also exhibits group sparsity.Thus,a method of TWRI exploiting multipath based on group sparse compressive sensing is proposed in this thesis.The main work of this thesis is as follows:(1)Brief introduction of the theoretical framework of TWRI based on group sparse CS.Based on the basic theories and principles of TWRI and CS,group sparsity is found on through-the-wall radar received signal model.Hence,a TWRI method based on group sparsity is proposed in this thesis.The performance of the proposed method is verified by two group sparse reconstruction algorithms respectively named SpaRSA and BOMP.(2)Development of BCS Based Group Sparse TWRI Technique.To exploit the group sparsity of TWRI,this thesis applies BCS to group sparse TWRI.For group sparse TWRI,complex multi-task BCS(CMT-BCS)algorithm is utilized to iteratively update Bayesian model parameters and reconstruct the image.It is shown in the simulation results that TWRI based on BCS can accurately and efficiently reconstruct the scene image behind a wall.(3)A TWRI method exploiting multipath based on group sparse BCS is proposed in this thesis.Group sparsity is also found on through-the-wall multipath based on further studies on thro ugh-the-wall radar multipath propagation modes.Hence,it can not only suppress the ghosts generated by multipath,but also improve the imaging performance of the through-the-wall radar by exploiting the group sparsity of multipath.Simulation results show that TWRI exploiting multipath based on group sparse CS can still accurately reconstruct targets in the scene while effectively suppressing ghosts and achieves high-quality imaging results in a through-wall multipath environment.