一种4D毫米波雷达高分辨角度估计方法

针对4D毫米波雷达中雷达孔径受限导致的角度分辨率不高的问题，提出一种改进的多重信号分类(Multiple Signal Classification,MUSIC)算法。首先通过数据预处理手段，补偿多输入输出(Multiple Input Multiple Output,MIMO)虚拟阵列中目标运动引起的相位误差;接着划分子阵并对角度维数据构成的协方差矩阵进行Toeplitz化处理，利用MUSIC算法估计一维角度;最后采用特征值排序的方法将估计出的俯仰角与方位角进行快速配对。该方法有效提高了4D毫米波雷达角度维估计的分辨率和估计精度，也提高了二维角度估计的计算速度。数值仿真与实物实验验证了所提方法在角度高分辨估计上的性能。

A High Resolution Angle Estimation Method for 4D Millimeter Wave Radar

In order to solve the issue of low angle resolution in 4D millimeter wave radar resulting from a restricted radar aperture,an enhanced multiple signal classification (MUSIC) algorithm is proposed.Firstly,the phase error caused by the movement of the target in the multiple input multiple output (MIMO) virtual array is compensated by the data pre-processing method.Then,the subarray is divided into three groups,the covariance matrix composed of diagonal-dimensional data is Toeplitz processed,and the MUSIC algorithm is used to estimate the one-dimensional angle.Finally,the estimated pitch angle and azimuth angle are quickly paired using the method of feature value sorting.This method effectively improves the resolution and estimation accuracy of 4D millimeter wave radar angle dimension estimation,and improves the computational speed of two-dimensional angle estimation.Numerical simulation and physical experiments verify the performance of the proposed method in high-resolution angle estimation.