Scribd
Upload
176 views4 pages

Wiener Hop F

This document discusses Wiener Hopf theory, which provides a method for estimating one signal from another in real time without delay. The theory involves factorizing the spectrum of the combined signals into parts with only positive or negative imaginary components. The cross spectrum between the input and output signals is divided by the negative imaginary component and a causal filter is constructed. This filter provides the best estimate of the input signal from the output signal and minimizes the mean square error between the actual and estimated input.

Uploaded by

Cristian Enrique Troncoso Garay
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
176 views4 pages

Wiener Hop F

This document discusses Wiener Hopf theory, which provides a method for estimating one signal from another in real time without delay. The theory involves factorizing the spectrum of the combined signals into parts with only positive or negative imaginary components. The cross spectrum between the input and output signals is divided by the negative imaginary component and a causal filter is constructed. This filter provides the best estimate of the input signal from the output signal and minimizes the mean square error between the actual and estimated input.

Uploaded by

Cristian Enrique Troncoso Garay
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

EE301a Professor: Lloyd R.

Welch Wiener Hopf Theory


1 Introduction

This theory arises when there are two related signals and it is desired to estimate the one signal from the other in real time", that is, without delay. The following gure illustrates a typical application.
N (t)

x(t)

H (jw ) c

y(t)

z(t)

desi gn fi l ter

v(t)

The tools used to design the lter involve the spectrum" and cross spectrum" of various signals. We begin with the spectrum of xt , Sxxj! , the channel frequency response, HC j!  and the spectrum of the interference, SNN j! . It is desired to nd a stable, causal lter applied to z t whose output, v t matches xt as closely as possible in the sense of minimizing
1

AVEfxt , v t2 g. The various properties of spectra and cross-spectra are:


1 2 3 4

Saaj! is real and  0


If at is real for all t then Saa,j!  Saaj! 
 j!  Sabj! = Sba

if bt = at  hC t then Sbaj! = SaaHC j!  j!  Sabj! = SxxHC Sbbj! = Saaj! jHC j!j2 if ct = at + bt then

Sccj! = Saaj! + Sabj! + Sbaj! + Sbbj!


For signals which are independent" the cross spectrum is zero. In the above gure we assume the signal xt and the interference, N t are independent so that SxN j! 0.
2

Wiener Hopf Theory

The explanation as to why the computation below gives the best estimate of the input at time t, given the output at all times f tg depends upon a knowledge of stochastic processes EE562a. Since EE562a is not a prerequisite for EE301a, we will have to accept it on faith.
3 Wiener-Hopf Calculation

This description applies only to the case when all spectra and cross-spectra are rational. That is, when the fromula for the spectra are ratios of polynomials in ! . Refering to the gure and the labels of the various signals the steps in the calulation: Step 1: Factor

Szz ! = G !  G,!


+

where G+ contains all of the zeros and poles with positive imaginary part in upper half plain and G, contains all of the zeros and poles with negative imaginary part in lower half plain.
3

Step 2: Divide Sxz by G, and do partial fraction expansion. 3 2 Sxz ! 7 X An 6 5= 4 G,! ! , an Step 3: Construct a causal transfer function by taking only those terms whose poles have positive imaginary part. 2 3 S  !  An X xz 7 6 4 5 = G,! + I man 0 ! , an Step 4: Divide by G+ to obtain the desired estimation lter: 2 3 1 6 S  !  7 Hdesign! = G+!  4 Gxz ,!  5 + Step 5: Rewrite in rational form ratio of two polynomials: !  Hdesign! = P Q!

You might also like