Research On SAR Imaging Method Of Automotive Millimeter Wave Radar

As a key component of the automotive intelligent driving assistance system,the automotive millimeter wave radar can detect the location of target objects in the scene by transmitting millimeter waves from the radar.However,there are still some problems with in vehicle millimeter wave SAR imaging: it is necessary to obtain high-precision velocity information of the vehicle platform and the spatial position of antenna sampling points;the limited accumulation of synthetic aperture in automotive millimeter wave SAR will lead to insufficient imaging azimuth resolution,and it is more susceptible to the influence of mirror reflection effects;existing 3D imaging methods are difficult to achieve high resolution while ensuring imaging efficiency.This article will focus on the above issues and conduct research on automotive millimeter wave radar SAR imaging methods from three aspects: platform velocity estimation,azimuth high-resolution focused imaging,and three-dimensional efficient high-resolution imaging.The specific work arrangement of this article is as follows:1.Aiming at the problem that millimeter wave SAR imaging requires platform velocity and spatial position of sampling points,a platform motion velocity estimation method that utilizes the signal correlation between MIMO radar channels is proposed.Firstly,the characteristics and correlation of echo signals from different transmission and reception channels under MIMO-SAR were analyzed.Based on this,the relationship between the correlation between the two channel echo signals and the platform motion distance was derived.Then,the delay that maximizes the correlation between the selected two sets of transmission and reception channel echo signals was extracted,and the platform velocity of each sampling point was calculated by combining the spatial positions of these two sets of transmission and reception channels,In order to reduce the velocity estimation error caused by few environmental targets or some extended targets parallel to the trajectory,the average instantaneous velocity of each sampling point estimated in a short period is used as the final estimated velocity.Finally,the spatial position of the sampling points can be obtained by further integrating the velocity.Under simulation experimental conditions,the average speed estimation error of this algorithm can reach within 3%;2.Aiming at the problems of low azimuth resolution and mirror reflection effect masking weak scattering targets in existing imaging methods,an extended target sparse focusing imaging method is proposed,which can effectively utilize the structural characteristics of targets in the azimuth direction in actual scenes and achieve weak scattering target focusing.This method first compresses the FMCW SAR echo data after motion compensation to obtain a one-dimensional range profile,and then extracts the range units with targets through one-dimensional CFAR.Only the range units with targets are imaged in azimuth using the proposed block fast iterative shrinkage thresholding algorithm(BFISTA)to solve the block sparsity problem for the range units with targets,Finally,the suppression of specular reflection and focusing imaging of weak scattering targets are achieved through the fusion of neighboring frame images.Under simulation experimental conditions,this algorithm can effectively distinguish two targets with an interval of 1.5°,and its effectiveness in real scenarios has been verified through measured data processing results;3.Aiming at the problem of existing automotive millimeter wave SAR 3D imaging methods are difficult to achieve high resolution while ensuring imaging efficiency,a three-dimensional sparse imaging algorithm based on FFT operator for MIMO-SAR is proposed.This algorithm utilizes the FFT operator and IFFT operator to establish a mapping between the scene image and the echo signal,and on this basis,a three-dimensional sparse imaging model is established to avoid the construction of the observation matrix.Then,the sparsity problem is solved using the FISTA to achieve three-dimensional high-resolution imaging,effectively reducing computational and storage costs.Similarly,the fusion of neighboring frame images is used to suppress specular reflection and focus imaging of weak scattering targets.Under simulation experimental conditions,this algorithm can effectively distinguish two targets with a height interval of about 6° in the vertical direction,and the effectiveness of this algorithm in real scenarios has been verified through measured data processing results.