Research On Key Technologies Of Ultra-wideband FMCW Radar Imaging

With the special advantages that traditional pulse radar does not have,Frequency Modulated Continuous Wave radar is able to achieve no-blind-area and high-resolution detection via low-cost,compact system and ultra-low sampling frequency.Accompanying with the improvement of hardware performance and theoretical level,FMCW radar has seen unprecedented development these years,and reflected tremendous applicable prospects both in military and civil use.This paper sticks closely to the subject of small air targets'high-resolution imaging,and endeavors to enhance the ability of target recognition.A wideband FMCW radar system working at X-band is designed.Taking this as an opportunity,the correction of nonlinear frequency modulation,Inverse Synthetic Aperture Radar?ISAR?imaging with a large rotation angle and interferometric ISAR three-dimensional imaging are researched in-depth.The main work and achievements of this paper are summarized as follows.1.Range measurement principles of triangular FMCW radar is researched,after which an X-band FMCW radar with a bandwidth of 5GHz is displayed.The framework,design principle,working process and performance index of the radar system are elaborated.Some crucial problems in wideband FMCW radar high-resolution imaging are analyzed on the basis of this specific X-band radar system.2.The problem of nonlinear modulation deteriorating range imaging in wideband FMCW radar is studied.The polynomial phase signal is employed to fit the nonlinear modulation,of which the characteristics and affection on range imaging are expounded.Then,two correction methods with clear pertinence are proposed in this paper.To solve the problem of error propagation in high-order ambiguity function algorithm and improve calculating efficiency,a nonlinearity correction method based on polynomial regression is firstly proposed.The polynomial coefficients can be estimated jointly with recursive least squares and phase unwrapping,which leads to an efficient and accurate nonlinearity correction result.To improve the anti-noise performance of estimation,a nonlinearity correction method based on Match Fourier Transform?MFT?is proposed.With image entropy and MFT,the optimal parameters can be obtained.Even in low SNR,sufficiently good performance of nonlinearity estimation and correction can be kept.Simulation and real data processing results demonstrate that both methods can achieve high-resolution range image.3.To obtain a high resolution in azimuth,the problem of ISAR imaging with a large rotation angle is researched.In the turntable model,the serious range-azimuth coupling of Range-Doppler?RD?algorithm is analyzed,followed by the Back Projection?BP?algorithm that can achieve well focusing performance.As the algorithms based on match filtering cannot suppress high sidelobes in ISAR image,the theory of Compressed Sensing?CS?is introduced in this paper.Then,with the help of blocked structure information of scatterers under high resolution,the method of ISAR imaging with a large rotation angle via adaptive block-sparsity matching pursuit is proposed.Through coherent projection,backtracking and adaptive update,the optimal atoms can be obtained.At last,an ISAR image is reconstructed without knowing the block sparsity.Simulation and real data processing results testify the feasibility and superiority of the proposed algorithm.4.To further improve target recognition capability,the theory of three-dimensional ISAR imaging is researched.Under the assumption that scatterers have the same coordinates in far-field multi-channel detection,the model of multiple measurement vectors?MMV?is introduced.To improve sparsity constraint of convex function penalty,a three-dimensional ISAR imaging method via joint sparsity based on non-convex optimization is proposed in this paper.The non-convex function is transformed to weighted l₁ norm,which can be solved via joint sparsity methods.Then the ISAR images are obtained and the scatterers 3-D coordinates are estimated.Simulation experiments show the effectiveness and the superiority of the proposed method.