毫米波FMCW MIMO雷达三维点云成像方法

毫米波FMCW MIMO雷达的三维点云成像在自动驾驶、智慧交通、工业和安防等领域的三维环境感知因其独特的优势而受到广泛关注。本文在现有处理框架下提出一种结合渐近最小方差稀疏迭代（SAMV）的高分辨点云成像方法。该方法首先对差拍信号做一维快速傅里叶变换（FFT）获得高分辨距离像，然后对每一个距离单元采用SAMV算法来估计方位角网格和俯仰角网格的功率，最后结合CASO-CFAR检测算法，得到探测区域内目标高分辨的距离、方位角和俯仰角三维信息。仿真和实验结果表明，相比现有结合Capon测角的方法，该方法能在相同雷达条件下角分辨率提高2倍，实现了毫米波雷达高分辨的三维成像。

3D Point Cloud Imaging Method with FMCW MIMO Millimeter Wave Radar

Three-dimensional (3D) environment perception technology is indispensable in communities such as autonomous driving, smart transportation, industry, and security. The 3D point cloud imaging of millimeter wave FMCW MIMO radar has received extensive research attention due to its unique advantages. This paper combines the asymptotic minimum variance sparse iteration (SAMV) under the existing signal processing framework to propose a high-resolution point cloud imaging method. This method first performs one-dimensional (1D) fast Fourier transform (FFT) on the beat signal to obtain a 1D high-resolution range profile; then uses the SAMV algorithm in each range cell to estimate the power of azimuth and elevation grids. Finally, the CASO-CFAR algorithm is used to detect range, azimuth and elevation 3D information with high-resolution of the target in the detection area. Simulation and experimental results show that, compared with the existing method incorporating Capon angle measurement, this approach can increase the angular resolution by 2 times under the same radar configuration, and realizes high-resolution 3D imaging of millimeter-wave radar.