NetNews Usenet Archive 1992 #27/NN_1992_27.iso/spool/sci/engr/control/292

Text File | 1992-11-18 | 1.7 KB | 44 lines

Newsgroups: sci.engr.control
Path: sparky!uunet!charon.amdahl.com!pacbell.com!ames!agate!stanford.edu!rock!taco!cmkurker
From: [e-mail] (CHRISTOPHER M KURKER)
Subject: continuous system identification
Message-ID: <[e-mail]>
Sender: [e-mail] (USENET News System)
Organization: NCSU Project Eos
Date: Wed, 18 Nov 1992 20:55:37 GMT
Lines: 33

Hello,

I hope this is an appropriate forum for a question about identification
of continuous-time (CT) and discrete-time (DT) systems.

In my research I need to parameterize a linear system that is the
cascade of a continuous-time system (an anti-aliasing filter with perhaps
two important poles and one important zero) followed by a discrete-time
system (a fourth-order DT Bessel filter).

My goal is to find the overall system's frequency response and various
measures of it, including 3dB frequency, phase linearity, etc.).

I need to parameterize the system from empirical data consisting of
frequency response data from the CT filter and time-domain impulse
response data from the DT filter.  (I suppose I could get time-domain
step response data for the CT filter also.)

I am aware of the various methods
for DT system identification for the DT filter (e.g. method of least
squares), but am unaware of what to do for the CT filter,
short of doing a DFT on the impulse response.  I am also aware
of direct frequency domain IIR DT system design methods which can calculate
a DT transfer function from magnitude and phase vs. frequency data,
and wonder if an analog of this exists for identifying CT systems.

Any help on this would be greatly appreciated.  Thanks.


Chris Kurker
[e-mail]
North Carolina State University
Raleigh, NC