TY - JOUR
T1 - Amplitude estimation of sinusoidal signals
T2 - Survey, new results, and an application
AU - Stoica, Petre
AU - Li, Hongbin
AU - Li, Jian
PY - 2000
Y1 - 2000
N2 - This paper considers the problem of amplitude estimation of sinusoidal signals from observations corrupted by colored noise. A relatively large number of amplitude estimators, which encompass least squares (LS) and weighted least squares (WLS) methods, are described. Additionally, filterbank approaches, which are widely used for spectral analysis, are extended to amplitude estimation. More exactly, we consider the recently introduced matched-filterbank (MAFI) approach and show that by appropriately designing the prefilters, the MAFI approach to amplitude estimation includes the WLS approach. The amplitude estimation techniques discussed in this paper do not model the observation noise, and yet, they are all asymptotically statistically efficient. It is, however, their different finite-sample properties that are of particular interest to this study. Numerical examples are provided to illustrate the differences among the various amplitude estimators. Although amplitude estimation applications are numerous, we focus herein on the problem of system identification using sinusoidal probing signals for which we provide a computationally simple and statistically accurate solution.
AB - This paper considers the problem of amplitude estimation of sinusoidal signals from observations corrupted by colored noise. A relatively large number of amplitude estimators, which encompass least squares (LS) and weighted least squares (WLS) methods, are described. Additionally, filterbank approaches, which are widely used for spectral analysis, are extended to amplitude estimation. More exactly, we consider the recently introduced matched-filterbank (MAFI) approach and show that by appropriately designing the prefilters, the MAFI approach to amplitude estimation includes the WLS approach. The amplitude estimation techniques discussed in this paper do not model the observation noise, and yet, they are all asymptotically statistically efficient. It is, however, their different finite-sample properties that are of particular interest to this study. Numerical examples are provided to illustrate the differences among the various amplitude estimators. Although amplitude estimation applications are numerous, we focus herein on the problem of system identification using sinusoidal probing signals for which we provide a computationally simple and statistically accurate solution.
UR - http://www.scopus.com/inward/record.url?scp=0033878985&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=0033878985&partnerID=8YFLogxK
U2 - 10.1109/78.823962
DO - 10.1109/78.823962
M3 - Article
AN - SCOPUS:0033878985
SN - 1053-587X
VL - 48
SP - 338
EP - 352
JO - IEEE Transactions on Signal Processing
JF - IEEE Transactions on Signal Processing
IS - 2
ER -