ASYMPTOTIC VARIANCE OF THE AR SPECTRAL ESTIMATOR FOR NOISY SINUSOIDAL DATA

被引：2

作者：

HANDEL, P

STOICA, P

SODERSTROM, T

机构：

[1] Systems and Control Group, Department of Technology, Uppsala University, S-75103 Uppsala

来源：

SIGNAL PROCESSING | 1994年 / 35卷 / 02期

基金：

瑞典研究理事会;

关键词：

AUTOREGRESSIVE MODELING; AR SPECTRAL ESTIMATOR; SPECTRAL LINE ANALYSIS; ASYMPTOTIC STATISTICAL ANALYSIS;

D O I：

10.1016/0165-1684(94)90041-8

中图分类号：

TM [电工技术]; TN [电子技术、通信技术];

学科分类号：

0808 ; 0809 ;

摘要：

In this paper the autoregressive (AR) spectral estimator is analyzed in the case of noisy sinusoidal data. An expression for the large-sample normalized variance is derived and studied in detail for increasing model order. In particular, a very simple formula is derived for the asymptotic (in both number of observed data and model order) normalized variance, which confirms a conjecture made by Sakai.

引用

页码：131 / 139

页数：9

共 8 条

[1] Angeby, Stoica, Soderstrom, Statistical analysis of autoregressive frequency estimates, Report UPTEC 91110R, (1991)
[2] Berk, Consistent autoregressive spectral estimates, The Annals of Statistics, 2, pp. 489-502, (1974)
[3] Gradshteyn, Ryzhik, Table of Integrals, Series, and Products, (1980)
[4] Kay, Modern Spectral Estimation, Theory and Application, (1987)
[5] Sakai, Statistical properties of AR spectral analysis, IEEE Trans. Acoust. Speech Signal Process., 27, pp. 402-409, (1979)
[6] Soderstrom, Stoica, System Identification, (1989)
[7] Stoica, Friedlander, Soderstrom, Asymptotic bias of the high-order autoregressive estimates of sinusoidal frequencies, Circuits Systems Signal Process., 6, 3, pp. 287-298, (1987)
[8] Stoica, Soderstrom, Ti, Overdetermined Yule-Walker estimation of the frequencies of multiple sinusoids: Accuracy aspects, Signal Processing, 16, 2, pp. 155-174, (1989)

← 1 →