PSN-L Email List Message

Subject: Re: Digest from 01/03/2013 00:01:02
From: David Sarraf david.sarraf@.........
Date: Fri, 4 Jan 2013 05:54:57 -0800 (PST)


Randy

The general concept is called a state observer.=A0 You construct a mathemat=
ical model of a system that runs in parallel with the physical system.=A0 I=
t will take the same inputs as the physical system and give an output that =
doesn't have all of the noise that the physical system produces.=A0=20

If you don't know the model parameters the state observer can generate them=
 for you.=A0 You make it adaptive by looking at inputs and outputs of the p=
hysical system and tuning the model parameters until you reduce the error, =
or the disagreement between the model and the system, to an acceptable valu=
e.=A0=20

Dave Sarraf
=A0

>I was reading some of the info on LIGO recently and it appears there is
>again use for some of this.=A0 As I understood the paper, the concept is t=
o
>run a simulation of the system behavior in parallel and compare the
>simulation to the real time measured response.=A0 The feed back loop is th=
en
>constructed as a smooth predicted value.=A0 This prevents amplification of
>noise developed in the actual system from feeding back into future amplifi=
Randy

The general concept is called a = state observer.  You construct a mathematical model of a system that r= uns in parallel with the physical system.  It will take the same input= s as the physical system and give an output that doesn't have all of the no= ise that the physical system produces. 

If you don't know the = model parameters the state observer can generate them for you.  You ma= ke it adaptive by looking at inputs and outputs of the physical system and = tuning the model parameters until you reduce the error, or the disagreement= between the model and the system, to an acceptable value. 

Da= ve Sarraf
 

>I was reading some of the info on LIGO recen= tly and it appears there is
>again use for some of this.  As I u= nderstood the paper, the concept is to
>run a simulation of the syste= m behavior in parallel and compare the
>simulation to the real time me= asured response.  The feed back loop is then
>constructed as a s= mooth predicted value.  This prevents amplification of
>noise de= veloped in the actual system from feeding back into future amplified error.=
[ Top ] [ Back ] [ Home Page ]