论文部分内容阅读
一、引 言 在诸如雷达、声纳、语言分析等许多信号处理领域中,一个经常遇到的共同问题就是要用有限的采样数据估计信号的谱。虽然传统的Blackman-Tukey算法和周期图算法可用来解决这一问题,但由于实际场合下可用的采样长度往往很短,计算出的谱的频率分辨力较低,结果并不很理想。 近年来,一些所谓“高分辨力谱分析方法”(High resolution spectrum analysis meth-ods)如最大似然(ML)、自回归(AR)、自回归和移动平均(ARMA)以及线性预测(LP)等方法日益受到重视。其中一个主要原因是:它们在信噪比不很低的情况下能提供比传统方法高得多的频率分辨能力。
I. INTRODUCTION In many areas of signal processing such as radar, sonar, speech analysis, a common problem commonly encountered is the use of limited sample data to estimate the spectrum of a signal. Although the traditional Blackman-Tukey algorithm and the periodic graph algorithm can be used to solve this problem, the calculated spectrum has a low frequency resolution due to the short sampling length available in practice and the unsatisfactory result. In recent years, some so-called High resolution spectrum analysis meth ods such as ML, AR, ARMA and LP, Methods such as increasing attention. One of the main reasons for this is that they offer much higher frequency resolution than traditional methods at low signal-to-noise ratios.