针对毫米波多输入多输出系统中的混合波束赋形问题，提出一种基于迭代更新思想的混合波束赋形算法，该算法通过提高模拟波束赋形矩阵与误差矩阵的相关性，获得了较优的系统性能。首先所提算法通过不断迭代更新得到误差矩阵，然后对误差矩阵进行奇异值分解得到左奇异值矩阵，并以此获得模拟波束赋形矢量，最后根据得到的模拟矢量更新模拟波束赋形矩阵。仿真结果表明，所提算法的性能优于传统的正交匹配追踪混合波束赋形算法，尤其当数据流数与射频链路数相等且数量较少时，所提算法的性能明显优于正交匹配追踪算法，且更近似于纯数字波束赋形的系统性能。

To solve the problem of hybrid beamforming in millimeter wave Multiple-Input Multiple-Output (MIMO) system, a hybrid beamforming algorithm based on iterative updating is proposed. The algorithm improves the system performance by improving the correlation between the analog beamforming matrix and the error matrix. Firstly, the proposed algorithm updates the error matrix iteratively. Then, the singular value decomposition of the error matrix is used to obtain the singular value matrix and then an analog beamforming vector is obtained. Finally, the analog beamforming matrix is updated according to the analog beamforming vector. Compared with the traditional Orthogonal Matching Pursuit (OMP) hybrid beamforming algorithm, the performance of the proposed algorithm is better than OMP algorithm and close to that of fully digital beamforming algorithm, especially when the number of data streams and Radio Frequency (RF) link is similar and small.