Title :
Frequency domain analysis of a family of minimax signal estimators
Author :
Sherman, Peter J.
Author_Institution :
Sch. of Mech. Eng., Purdue Univ., West Lafayette, IN, USA
fDate :
3/1/1988 12:00:00 AM
Abstract :
A minimax estimator of the mean of x~N(θ, Rn) is evaluated, where Rn is Toeplitz. The problem can be viewed as estimating a signal θ that has been corrupted by additive wide-sense stationary noise. The evaluation is carried out using a frequency-domain approach that yields advantages in design and analysis. In particular, counterintuitive performance of an entire class of minimax forms is discovered when R n is not diagonal. A numerical example is presented to demonstrate the performance
Keywords :
estimation theory; frequency-domain analysis; minimax techniques; signal processing; additive wide-sense stationary noise; counterintuitive performance; frequency domain analysis; minimax signal estimators; Additive noise; Eigenvalues and eigenfunctions; Frequency domain analysis; Frequency estimation; Maximum likelihood estimation; Minimax techniques; Signal analysis; Signal processing; Working environment noise; Yield estimation;
Journal_Title :
Acoustics, Speech and Signal Processing, IEEE Transactions on
