论文部分内容阅读
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(covariance matching estimation technique,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:NCCOMET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“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.
In order to solve the problem of azimuth dependent amplitude and phase errors in non-circular signal direction finding, this paper presents a covariance matching estimation technique (COMET) based on the non-circular features of non-circular signals and the self-correction of auxiliary elements An improved algorithm is proposed for NCCOMET, which is applicable to both unknown source statistical properties in the time domain and a priori known statistical independent features: The algorithm uses non-circular signals to extend the covariance data to make the correction accuracy more conventional The maximum likelihood algorithm based on auxiliary elements (without the use of non-circular features) is obviously improved, and the minimum number of auxiliary elements is reduced.The statistical consistency of the parameter estimation is theoretically proved by the first-order error analysis The mean square error of the parameter estimation under the influence of finite sampling is given and the definition of “data utilization” of the algorithm is proposed. The quantitative comparison of the data utilization rate of the NC-COMET algorithm with that of the conventional maximum likelihood algorithm The simulation results also show that the performance of NC-COMET algorithm is better than the conventional maximum likelihood algorithm: robustness under low signal-to-noise ratio In the case of a priori known or large non-circularity of the eigenvalue, the algorithm is particularly effective in improving the accuracy of the calibration.