本文提出了改进的秩-相位误差(IROPE)估计自聚焦算法。与ROPE算法相比,IROPE有两个优点:(1)它基于的模型更符合实际,所以具有更好的鲁棒性。(2)它对迭代过程初值的设定是近似准确的。本文通过近似分析证明,IROPE对相位误差差分的估计是最大似然(ML)估计,而相位梯度自聚焦(PGA)算法是IROPE的一个子集,即在满足一定条件时,IROPE和PGA对相位误差的估计都是线性无偏最小方差估计(LUMV)。 This paper presents an improved self-focusing algorithm for rank-phase error estimation (IROPE). Compared with the ROPE algorithm, IROPE has two advantages: (1) It is based on the model more realistic and therefore has better robustness. (2) It is approximate to the initial value of the iterative process. In this paper, approximate analysis shows that IROPE estimates the phase error difference as the maximum likelihood (ML) estimation, while the phase gradient auto-focusing (PGA) algorithm is a subset of IROPE. That is, when certain conditions are met, The estimates of the errors are all Linear Unbiased Minimum Variances Estimates (LUMVs).