一种快速的谱分析迭代法  

A Fast Iterative Algorithm Based on Spectral Analysis

作　　者：马建林[1] 张娜敏 郭汝江[1] 

机构地区：[1]南京电子技术研究所,南京210039 [2]南京恩瑞特实业有限公司,南京211106

出　　处：《现代雷达》2012年第3期30-34,共5页Modern Radar

摘　　要：为克服现有正弦波频率估计算法运算量大或估计精度不高的弊端,提出了一种快速的频率估计算法,该算法利用正弦波频谱对称性原理和频移技术,使信号频率位于峰值点附近再进行估计。仿真结果表明:文中算法稳定性不随被估计频率分布而产生波动,在低信噪比条件下不存在发散现象,以它为初始值进行一次谱分析迭代,其均方根误差在整个频段都接近克拉美-罗限。Traditional sinusoid frequency estimation methods are commonly suffered by high computing costs or poor estimation pre- cision. To resolve these issues, a novel fast frequency estimation algorithm is proposed. Utilizing the spectrum symmetry principle of sine wave and frequency shift technology, this method can shift the signal frequency near the peak before re-estimating the fre- quency. Simulation results indicate that the stability is robust with the frequency distribution, and has no divergence phenomenon in the low SNR condition. Further investigation is operated by spectral analysis with only one iterating, in which the result of the new algorithm is set as initial value. The root mean squared error （RMSE） of the spectral analysis approaches the Cramer-rao lower bound （CRLB） over the whole frequency band.

关 键 词：频率估计 辛格函数 对称性 谱分析 克拉美-罗限 

分 类 号：TN957.51[电子电信—信号与信息处理]