有限长观测数据的最大熵谱分析与算法

利用长度有限的观测数据估计随机过程的功率谱，是随机信号理论的主要应用之一，常规的谱分析方法，往往无法兼顾分辨力和平滑性两方面的需要，而用最大熵谱分析方法则可得到分辨力很高的，平滑的谱估计，在观测数据较少时，这种方法更能显示出其较之于常规谱分析方法的优越性，本文简要介绍了最大熵谱分析方法的基本概念及算法，通过计算机模拟的结果和常规的谱分析方法进行了比较和分析，从中得出了一些较有实用价值的结论。

The Maximum Entropy Spectral Analysis and Its Algorithm on Finite Length Observed Data

Estimating the power spectrum of stochastic processes with finite lengthobserred data is one of the primary applications of random signals theory. It can't be satis-fied for the requirement of resolution and smoothness simultaneously with the conventionalspectrum analysis methods,but use maximum entropy spectral analysis,it can get a spec-trum estimation with high-resolution and smoothness.If there is a little observed data,itwill shows that this method to have some advantages over the conventional methods.Thispaper gives an introduction to the fundamental conceptions of maximuin entropy spectralanalysis and its algorithm to implement the estimation of random signals'powerspectrum.Some computerized simulative results were analysed and compared, and wehave come to some practical conclusions.