非圆信号的协方差匹配自校正算法及其性能分析

针对非圆信号测向中方位依赖幅相误差的校正问题，本文根据非圆信号的非圆特性和辅助阵元能够自校正的特点，对协方差匹配估计技术（covariancematchingestimationtechnique，COMET）进行改进，提出一种适用于信源时域统计特性未知和统计独立特性先验已知两种情况的改进算法：NC—COMET算法．该算法利用非圆信号扩展协方差数据，使其校正精度较常规的基于辅助阵元的最大似然类算法（未利用非圆特性）有明显提升，且降低了最小辅助阵元数要求．从理论上证明了参数估计的统计一致性，采用一阶误差分析方法推导了有限采样影响下参数估计的均方误差表达式，并提出算法的cc数据利用率”定义，定量比较获得了NC．COMET算法的数据利用率较常规的最大似然类算法的提升幅度．仿真结果亦表明NC—COMET算法性能较常规的最大似然类算法更优：低信噪比下具有更强的鲁棒性；信源时域统计独立特性先验已知或者大非圆率的情况下，该算法对校正精度的提升尤为明显．

Auto-calibration method based on covariance matching estima- tion for noncircular sources and its performance analysis

An improved direction-finding algorithm for noncircular signals was proposed in the presence of an- gularly dependent gain and phase errors, which is called NC-COMET based on the fundamental principle of covariance matching estimation technique （COMET） by using instrumental sensors for achieving autocalibration. The proposed algorithm is effective in the absence of time-domain statistical information as well as in the pres- ence of statistical independent information of sources. By applying the data of extended covariance matrix of noncircular signals, its calibration accuracy is improved significantly compared with conventional maximum like- lihood Mgorithm which does not take advantage of noncircular feature, and the minimum number of the needed instrumental sensors is decreased. The performance study on NC-COMET proved the statistical consistency of the parameter estimation, and a theoretical derivation for the closed-form expression of the mean square error （MSE） of NC-COMET estimation was presented under the influence of finite samples by the first-order analysis. In addition the definition of ＂signal utilization rate＂ was given and the signal utilization rate of NC-COMET algorithm was certified to be higher than that of the conventional maximum likelihood algorithm by quantita- tive comparison. The simulation results also showed the performance advantages of NC-COMET algorithm over conventional maximum likelihood algorithm. The proposed algorithm is more robust with respect to low signal- to-noise ratio, and its estimation is much more precise in the presence of statistical independent information of sources or when the sources have larger circularity rate.