Research On Forward-looking Imaging Algorithms Based On Millimeter Wave Radar

Forward-looking imaging radar can achieve range-azimuth two-dimensional imaging of the forward scene of the moving platform.It has great application value in autonomous landing,precision guidance and other fields,and has become an international research hotspot.High range resolution imaging can be obtained by transmitting broadband LFM signals and using pulse compression technology.In azimuth direction,because the direction of motion of the platform is almost perpendicular to the azimuth direction of the imaging scene,the forward-looking radar can not accumulate large Doppler bandwidth,so it is difficult to achieve high azimuth resolution imaging with synthetic aperture.In addition,due to the limited space size of the moving platform,it is difficult for traditional centimeter and decimeter wave radar to form narrow beams,and the beam sharpening technology alone can not meet the demand of modern radar for high azimuth resolution.Because of its small wavelength,millimeter wave radar can form a narrow beam in a small size space.On this basis,azimuth super-resolution imaging will become an important means to improve the azimuth resolution of forward-looking imaging radar.In this master thesis,the forward-looking imaging technology of millimeter wave radar is investigated.The theoretical research of forward-looking imaging algorithm based on monopulse angle measurement,regularized deconvolution and Bayesian deconvolution is carried out.The feasibility of each imaging algorithm is verified by simulation.The main contents and innovations are as follows:1.According to the geometric relationship between radar and target,the echo signal model of millimeter wave forward-looking imaging radar is established.On this basis,the basic imaging process and method of millimeter wave forward-looking imaging radar are proposed by analyzing the echo characteristics of range and azimuth and combining the working principle of forward-looking imaging radar.2.On the basis of monopulse angle measurement,the imaging model of monopulse angle measurement is established by calculation and theoretical deduction,and a millimeter wave forward-looking imaging method based on monopulse angle measurement is realized.The simulation of point target and scene target shows that the algorithm is effective for accurate measurement and imaging of target angle.3.The azimuth deconvolution model of millimeter wave forward-looking imaging radar is established.L1-TSVD regularization method is proposed for sparse target imaging.Compared with L1 regularization imaging algorithm,this algorithm has the advantages of strong noise suppression and better robustness.L2 regularization method is used for scene imaging,which can effectively suppress the noise on image.4.The forward-looking radar imaging is realized in the Bayesian framework,and SBL(Sparse Bayesian Learning)is introduced into the forward-looking imaging radar.SBL-TSVD algorithm is proposed to image sparse scenes.Compared with the L1 regularization imaging algorithm,this method can make full use of the echo data and automatically derive the Super-parameters from it,and the imaging effect is better.