DocumentCode :
745705
Title :
An Algebraic Approach to the Estimation of the Order of FIR Filters From Complete and Partial Magnitude and Phase Specifications
Author :
Constantinides, Anthony G. ; Li, Wen Min
Author_Institution :
Dept. of Electr. & Electron. Eng., Imperial Coll. London
Volume :
55
Issue :
4
fYear :
2007
fDate :
4/1/2007 12:00:00 AM
Firstpage :
1213
Lastpage :
1222
Abstract :
The problems addressed by this paper is the following: Given a set of measurements over the range of normalized frequencies (thetas1 ,thetas2) on the magnitude and/or phase of a real FIR but otherwise unknown filter, to estimate the order of the FIR filter. The range (thetas1,thetas2) may be partial or it may cover the entire range of frequencies. The purpose of the paper is to propose a new algebraic approach to solve the above collection of problems. Specific new results include FIR order estimation from partial or complete noiseless measurements for a real system from: a) phase alone, from b) magnitude alone (not necessarily piece-wise constant), and from c) joint magnitude and phase. The proposed approach is not only capable of dealing with specifications that go beyond the conventional formulas for the standard piece-wise-constant magnitude FIR filter order estimation, but it also furnishes a nexus for order estimation from phase (or group delay) specifications, areas which have remained hitherto unaddressed. The approach is based on the use of Root Moments. In this context, the novel concept of Fractional Root Moments is used in a key fashion to provide an estimate on the number of zeros inside the unit circle. Open problems and new directions of exploration and research are mentioned in the body of the paper
Keywords :
FIR filters; Newton method; FIR filters; fractional root moments; magnitude specifications; noiseless measurements; phase specifications; piece-wire-constant magnitude; Delay estimation; Digital filters; Finite impulse response filter; Frequency estimation; Frequency measurement; Noise measurement; Phase estimation; Phase measurement; Phase noise; Signal processing algorithms; FIR filters; Newton identities; fundamental relationships; root moments;
fLanguage :
English
Journal_Title :
Signal Processing, IEEE Transactions on
Publisher :
ieee
ISSN :
1053-587X
Type :
jour
DOI :
10.1109/TSP.2006.887565
Filename :
4133022
Link To Document :
بازگشت