校正源方位存在偏差时的幅相误差顽健校正算法

摘要

The direction deviations of the calibration sources would significantly degrade the calibration accuracy of array gain-and-phase errors. Aiming to this problem,a robust calibration algorithm for gain-and-phase errors against the location deviations was presented under the assumption that the prior probability distribution of the location deviation was known. The idea of the robust algorithm was based on subspace fitting criterion and Bayesian estimation theory framework. The gain-and-phase errors were obtained through calculating the eigenvector associated with the minimum eigenvalue of some real symmetric matrix without estimating the locations of calibration sources. The closed-form Cramer-Rao bound(CRB) for the unknown parameters was derived,and the asymptotic distributions of the robust algorithm as well as the Cheng method were also given in the presence of direction deviations. Both the theory analysis and the simulation experiments validate that the asymptotic performance of the robust algorithm can reach the CRB under some moderate conditions and outperform the Cheng method when the location deviations exist.%针对校正源方位偏差会影响幅相误差校正精度这一问题,在假设校正源方位偏差的概率分布已知的条件下,依据子空间拟合准则和Bayesian估计理论框架,给出了一种抑制校正源方位偏差的幅相误差顽健校正算法.该算法可在无需估计校正源方位的情况下,通过计算某实对称矩阵最小特征值对应的特征向量获得幅相误差的顽健估计.推导了参数估计的CRB(Cramer-Rao bound),分析了顽健算法的渐近性能以及Cheng方法在校正源方位有偏差时的渐近性能.理论分析和仿真实验均表明:在一定条件下,所提出的算法的渐近性能可达到CRB,并且优于Cheng方法的渐近性能(当校正源方位有偏差时).