Help to simulate underwater environment

Hello 

  I have been working on a signal detection algorithm assessing its
performance for signals affected by AWGN noise. Now I have to test the
algorithm's efficacy in a underwater environment.
  What is the type of channel model to be employed?
  If possible, give details on how to simulate this channel, or
suggets some links/newsgroups where I can get more information on it?


Thanks in advance

Parthasarathy

Hello 

  I have been working on a signal detection algorithm assessing its
performance for signals affected by AWGN noise. Now I have to test the
algorithm's efficacy in a underwater environment.
  What is the type of channel model to be employed?
  If possible, give details on how to simulate this channel, or
suggets some links/newsgroups where I can get more information on it?


Thanks in advance

Parthasarathy

Hi Parthasarathy

Parthasarathy wrote:

>Hello 
>
>  I have been working on a signal detection algorithm assessing its
>performance for signals affected by AWGN noise. Now I have to test the
>algorithm's efficacy in a underwater environment.
>  What is the type of channel model to be employed?
>  If possible, give details on how to simulate this channel, or
>suggets some links/newsgroups where I can get more information on it?
>
>
>Thanks in advance
>
>Parthasarathy
>  
>
Would this signal be of the acoustic or radio variety?

Steve

parth175@yahoo.co.in (Parthasarathy) wrote in message news:<7f126353.0411092027.32ec5f9@posting.google.com>...
> Hello 
> 
>   I have been working on a signal detection algorithm assessing its
> performance for signals affected by AWGN noise. Now I have to test the
> algorithm's efficacy in a underwater environment.
>   What is the type of channel model to be employed?

What water depths are you interested in? Littoral waters? 
Deep-ocean basins?  What's the frequency range on your signal? 
Does your signal processing schme work in time domain or in 
frequency domain? What type of bathymeyry do you want to include? 
What types of sea floor geologies do you want to include? Would 
you like to include tidal effects? Weather effects? Oceanographic 
effects? Marine animals or vegetation? What geographical scales 
are you interested in? Basin-wide? Local?

Rune

Parthasarathy wrote:
> Hello 
> 
>   I have been working on a signal detection algorithm assessing its
> performance for signals affected by AWGN noise. Now I have to test the
> algorithm's efficacy in a underwater environment.
>   What is the type of channel model to be employed?
>   If possible, give details on how to simulate this channel, or
> suggets some links/newsgroups where I can get more information on it?
> 
> 
> Thanks in advance
> 
> Parthasarathy

Try

http://www.hlsresearch.com/oalib/

The models at oalib tend to be deterministic models.

If you have Matlab, I kind of like Prosim from the SACLANT center in 
Italy.  It includes some stochastic modeling features.

Most models don't include doppler effects.

> >
> Would this signal be of the acoustic or radio variety?
> 
> Steve


Hello Steve

  Acoustic signal

with regards
Parthasarathy

What is the type of channel model to be employed?
> 
> What water depths are you interested in? Littoral waters? 
> Deep-ocean basins?  What's the frequency range on your signal? 
> Does your signal processing schme work in time domain or in 
> frequency domain? What type of bathymeyry do you want to include? 
> What types of sea floor geologies do you want to include? Would 
> you like to include tidal effects? Weather effects? Oceanographic 
> effects? Marine animals or vegetation? What geographical scales 
> are you interested in? Basin-wide? Local?
> 
> Rune


Hello Rune

Goodness gracious, There are so many factors.
Please suggest a reference, to learn more about these. 
Thanks in advance.

Parthasarathy

Sonar performance prediction model 'ALMOST'

http://www.tno.nl/instit/fel/os/prod/almost/index.html

Ren=E9

parth175@yahoo.co.in (Parthasarathy) wrote
>   I have been working on a signal detection algorithm assessing its
> performance for signals affected by AWGN noise. Now I have to test the
> algorithm's efficacy in a underwater environment.
>   What is the type of channel model to be employed?
>   If possible, give details on how to simulate this channel, or
> suggets some links/newsgroups where I can get more information on it?

There is a nice "tutorial" article with illustrations and further
references in the October 2004 issue of Physics Today. (It focuses on
the shallow water scenario, but you might find it interesting
nonetheless.)

http://www.mpl.ucsd.edu/people/wak/papers/pt_Kuperman10_2004.pdf

Jeremy Furtek

parth175@yahoo.co.in (Parthasarathy) wrote in message news:<7f126353.0411112118.2eff784f@posting.google.com>...
> What is the type of channel model to be employed?
> > 
> > What water depths are you interested in? Littoral waters? 
> > Deep-ocean basins?  What's the frequency range on your signal? 
> > Does your signal processing schme work in time domain or in 
> > frequency domain? What type of bathymeyry do you want to include? 
> > What types of sea floor geologies do you want to include? Would 
> > you like to include tidal effects? Weather effects? Oceanographic 
> > effects? Marine animals or vegetation? What geographical scales 
> > are you interested in? Basin-wide? Local?
> > 
> > Rune
> 
> 
> Hello Rune
> 
> Goodness gracious, There are so many factors.

Well, that's what you need to deal with at sea. Depending on the 
scale of the experiment and exactly what you want to find out, 
different physical effects are important. Underwater acoustics
is 10% science and 90% voodoo. The voodoo part relates to seeing 
which effects are important and which are not, in any given experiment.

> Please suggest a reference, to learn more about these. 

Jensen & al: "Computational Ocean Acoustics"
             Springer, 2000.

Be a bit careful about the signal processing techniques those guys 
suggest, though. That Matched Field Processing stuff is not something 
you would want to get involved with.

If you just need something quick'n dirty up and running, try the PROSIM 
package Stan suggested. There are some ray tracing models available via 
that OALIB site as well, so you can get both frequency domain and time 
domain models.

Rune