Research On Frequency Diverse Array Imaging Based On Compressed Sensing

Downward-looking linear array three-dimension synthetic aperture radar(SAR)synthesizes a virtual two-dimensional array by controlling the movement of the antenna,combined with pulse compression technology,to obtain the three-dimensional resolution of scene targets.However,the hardware design of this kind of SAR system based on wideband transmission signals is complicated and the received signals are difficult to separate.By applying the frequency diverse array(FDA)to the 3D-SAR model,each array element only needs to transmit a single frequency signal to obtain broadband observation performance,which greatly reduces the hardware requirements of system.However,due to the space-frequency sparseness of FDA echo signals,the resolution is limited and the sidelobes of the radar images are relatively high when using the traditional algorithm based on matched filtering.The compressed sensing method transforms the echo signal into a linear programming model to reconstruct the spatial target informat ion and realize high-resolution imaging of the target.By using compressed sensing algorithm,FDA-3D-SAR can accurately reconstruct the information of space targets when the echo is sparse,and overcome the disadvantage of low accuracy of traditional imaging methods.Based on FDA-3D-SAR imaging system model,considering the hardware cost and memory occupancy constraints,this paper aims to achieve high-resolution and low sidelobes imaging of 3D scene targets by using compressed sensing method,and analyzes the characteristics of compressed sensing reconstruction algorithm applied to the FDA radar.The main work of this paper is as follows:(1)In view of the large amount of data to be processed by the FDA-3D-SAR system and the high sidelobes of traditional imaging algorithms,this paper proposes a random FDA-3D-SAR imaging method based on compressed sensing.The array elements in the tangent-track and the observation positions in the along-track are selected randomly and sparsely to realize two-dimensional sparse sampling of echo data.In the imaging part,orthogonal matching pursuit(OMP)algorithm and Bayesian compressed sensing(BCS)algorithm are used to reconstruct the scattering coefficient of targets.This paper also analyzes the characteristics and imaging performance of the two reconstruction algorithms applied to random FDA-3D-SAR system.Simulation and experimental results show that compressed sensing method not only reduces the data processing amount of FDA-3D-SAR system during imaging,but also effectively suppresses the sidelobes of radar images,and the imaging quality is significantly improved.(2)Aiming at the problem that FDA with random frequency offset is difficult to implement in hardware design,this paper proposes an improved method of selecting frequency offset coefficient of FDA array element based on Gray code.This method effectively decouples the beam information while constraining the degree of freedom of frequency increment.On this basis,combined with SAR technology using fixed frequency offset in the along-track,this paper proposes a FDA-3D-SAR system model based on Gray code,and also uses compressed sensing method to image the targets.The simulation results show that 3D radar image can be obtained clearly by using the compressed sensing reconstruction algorithm,which verifies the rationality of the imaging model and the practicability of the compressed sensing method.(3)In order to solve the problem that the dimensions of compressed sensing echo sensing matrix corresponding to FDA-3D-SAR imaging scene are too large,this paper proposes a method to reduce the number of columns of the echo sensing matrix based on coherent integration extraction preprocessing.Combined with the sparse sampling method of the number of rows of the echo sensing matrix,the dimensions of the echo sensing matrix can be greatly reduced while the spatial target information is accurately reconstructed,and the amount of imaging data calculation and system memory usage are further reduced.