摘要
周期图法是功率谱估计的一种基本方法,但该方法不满足一致性估计条件,谱估计的分辨力和方差都很难满足实际应用需要。因此采用其改进方法(即Welch法)估计信号功率谱,并结合该方法对四种典型窗函数的谱估计性能进行了具体的仿真,就其中的频率分辨率、采样信号的数据长度与窗函数的特性等因素之间的关系,进行了详细的讨论,重点分析了窗函数对谱估计的影响,指出了它们的优缺点,得出了不同数据类型与估计方式下的功率谱标准方差的均值。最后,根据实验结果并结合窗函数的评价指标,指出了选取合适的窗函数应注意的一般原则,具有一定的实际指导意义。
Periodogram is a basic method for power spectrum estimation, but it has too many deficiencies in resolution, variance and consistency constraint to meet the requirement well in practice. So this paper introduces its modified form （i. e. Welch method） to estimate the power spectrum of signals, and then discusses the relationships among the resolution, the length of sampling sequence and the characters of window function in power spectrum estimation based on Welch. Meanwhile, it primarily analyzes the influence of window functions and points out their virtues and defects. With the simulation of examples using MATLAB, the estimation performance of four different windows and the mean of standard deviation of power spectrum under different length and estimation methods have been figured out. Finally, some suggestions referring to experimental results have been presented when choosing windows, which are useful in fact.
出处
《计算机仿真》
CSCD
2008年第3期111-114,共4页
Computer Simulation
关键词
周期图
谱估计
窗函数
韦尔奇法
Periodogram
Spectrum estimation
Window function
Welch method
