论文部分内容阅读
引言尽管可以利用现代的计算设备,但现代的高性能DSP算法在实用系统中的应用仍受到很大限制。现今文献中提出的“高分解率”数据处理算法目前有多种变型,它们包括Burg提出的众所周知的自回归谱估计技术,Schmitt的“MUSIC”算法,Roy、Paulrai和Kai-1ath最近发表的“ESPRIT”算法,以及Nickel和Bǒhme建议的迭代最大似然法。未能直接应用的原因看来根源在于很多这种技术缺乏韧性(robustness)。如卡尔曼滤波器,在设计条件下,它们往往工作得很好,但是,在非理想情况下却易严重退化。本文的目的有三点。第一,讨论高分解率技术的原理和有关韧性问题的基本缘由。第
INTRODUCTION Although modern computing devices can be utilized, the use of modern, high-performance DSP algorithms in practical systems is still severely limited. There are currently many variations of the “high resolution” data processing algorithms proposed in the literature, including well-known autoregressive spectral estimation techniques proposed by Burg, Schmitt’s “MUSIC” algorithm, Roy, Paulrai and Kai-1ath’s recently published “ ESPRIT ”algorithm, and the iterative maximum likelihood method proposed by Nickel and Bǒhme. The reason for the failure to apply directly seems to lie in the fact that many of these technologies lack robustness. Such as Kalman filters, tend to work well under design conditions but can degrade badly under less than ideal conditions. The purpose of this article is three points. First, discuss the rationale for high resolution techniques and the basic rationale for toughness issues. No.