[kepler-dev] domain for signal procesing

[Edward A Lee]

At 03:57 PM 6/16/2004 -0700, Tobin Fricke wrote:
>Ptolemy's ``simple signal processing'' examples work with sampled data
>that have globally synchronous sampling points and a fixed, universal
>sampling rate; these examples therefore are well-suited to the SDF domain.
>
>I'm wondering what would be the best model of computation for a more `real
>world' signal processing system.  Signals could have idiosyncratic
>sampling rates and sampling that is in general not synchronized.  Here
>each token is a measured value of a signal at an associated time stamp,
>and relations have an associated `sample rate' property.  This sounds very
>much like the distributed discrete events (DDE) domain.

  <snip>


>thanks,
>Tobin Fricke


Asynchronous sampling can be handled using the PN domain (see
http://ptolemy.eecs.berkeley.edu/publications/papers/95/parksThesis/)
This thesis talks in some depth about a CD to DAT sample rate conversion
application... As you may know, the conversion here was deliberately
made technically difficult to discourage piracy.

Note that the FIR filter in Ptolemy II supports polyphase multirate
interpolation and decimation, so you can do quite sophisticated
sample rate conversions.

Asynchronous sample rates can also be handled using heterochronous dataflow 
(see
http://ptolemy.eecs.berkeley.edu/publications/papers/99/starcharts/
and
http://ptolemy.eecs.berkeley.edu/papers/03/communicationModeling/index.htm)
This technique hierarchically combines finite state machines and
synchronous dataflow graphs. It has been applied to timing recovery
in digital communication systems, where the sample rate is dynamically
determined by a PLL.

Parameterized synchronous dataflow (PSDF) can also deal with it (see
http://ptolemy.eecs.berkeley.edu/publications/papers/04/unifiedReconfigurationMemocode/
and the PSDF demo in the 4.0 Ptolemy II release). The PSDF demo in
4.0 includes the timing recovery problem as part of blind estimation
of a communication signal with unknown modulation schemes.

Of course, you can also jointly model continuous time signals using
CT and you can sample them at arbitrary times... This requires, of course,
having an accurate model of the continuous-time signals, which not
all applications have.  Typically this makes sense if you have a model
of the physical phenomenon that is producing the signal.

Hope this helps...
Edward



------------
Edward A. Lee, Professor
518 Cory Hall, UC Berkeley, Berkeley, CA 94720
phone: 510-642-0455, fax: 510-642-2739
eal at eecs.Berkeley.EDU, http://ptolemy.eecs.berkeley.edu/~eal