非高斯ARMA噪声中谐波恢复的杂交ESPRIT方法  

THE HYBRID ESPRIT APPROACH TO HARMONIC RETRIEVAL IN NON-GAUSSIAN ARMA NOISE

作　　者：梁应敞[1] 戴逸松[2] 王树勋[2] 

机构地区：[1]清华大学自动化系,北京100084 [2]吉林工业大学电子工程系,长春130025

出　　处：《电子科学学刊》1994年第6期561-568,共8页

摘　　要：本文研究非高斯ARMA噪声中的谐波恢复问题,提出了一种基于二阶和三阶统计量的杂交ESPRIT方法,该方法先估计噪声过程的AR部分参数,然后对观测值进行预滤波,最后估计谐波信号参量。模拟实验还验证了该方法的有效性和高分辨率。This paper addresses the harmonic retrieval problem in non-Gaussian ARMA noise. A hybrid ESPRIT approach using second- and third-order statistics is proposed. First, third-order statistics are used to identify the AR part of the non-Gaussian noise process, then the noisy measurements are filtered by AR polynomial, finally, the harmonic signal parameters are estimated. Simulation examples show the effectiveness and high resolution of the new approach.

关 键 词：非高斯 ARMA 噪声 谐波恢复 杂交 ESPRIT方法 

分 类 号：TN911.4[电子电信—通信与信息系统]