| 摘要: |
| 针对现有的车载毫米波雷达高分辨成像方法因存在离网误差导致精度低的问题，通过利用最大似然（ML）准则的统计特性，提出了一种迭代最小化稀疏学习（SLIM）与ML估计相结合的高精度测角方法用于点云成像。首先对中频数据进行距离维FFT得到距离像，再对新的数据进行多普勒维FFT得到多普勒像，经恒虚警率检测（CFAR）之后获取目标的距离、速度索引点，然后对索引点对应的多个虚拟接收阵元数据使用SLIM方法进行角度估计，最后通过最小化ML成本函数来细化波达方向（DOA），从而获得高精度的点云像。仿真和实验结果表明，该方法得到的点云像具有高精度的特点，角度精度能够达到0.1°。 |
| 关键词: 车载毫米波雷达 离网误差 最大似然 迭代最小化稀疏学习 高精度成像 |
| DOI：DOI:10.3969/j.issn.1672-2337.2023.05.016 |
| 分类号:TN957.52 |
| 基金项目:广西创新驱动发展专项（No.桂科 AA21077008）；广西无线宽带通信与信号处理重点实验室 2023年主任基金；桂林电子科技大学研究生教育创新计划项目（No.2022YCXS028） |
|
| A High Precision Imaging Method for Vehicular Millimeter Wave TDM⁃MIMO Radar |
|
MO Liangsheng， JIN Liangnian
|
|
1. School of Information and Communication, Guilin University of Electronic Technology, Guilin 541004, China;2. Guangxi Key Laboratory of Wireless Broadband Communication and Signal Processing, Guilin 541004, China
|
| Abstract: |
| Aiming at the problem that the existing high?resolution imaging methods for vehicle?mounted millimeter wave radar have low accuracy due to off?grid errors, a high?precision angle?measuring method combining iterative minimum sparse learning（SLIM） and maximum likelihood （ML） estimation is proposed for point cloud imaging by using the statistical characteristics of ML criterion. First, the range profile is obtained by range dimension FFT on intermediate frequency data, and then the Doppler image is obtained by Doppler dimension FFT on new data. After the constant false alarm rate（CFAR） detection, the range and velocity index points of the target are obtained. Then the angle estimation of the multiple virtual receiving array metadata corresponding to the index points is performed using the SLIM method. Finally, the direction of arrival（DOA） is refined by minimizing the ML cost function to obtain high?precision point cloud images. The simulation and experimental results show that the point cloud images obtained by this method has high accuracy, and the angle accuracy can reach 0.1°. |
| Key words: vehicular millimeter wave radar off⁃grid error maximum likelihood sparse learning via iterative minimization high precision imaging |
