Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Register Log in

higher order moments estimation in nongaussian noise

Status
Not open for further replies.
Basil

Basil

Member level 5
Joined
Jan 14, 2003
Messages
93
Helped
2
Reputation
4
Reaction score
0
Trophy points
1,286
Activity points
894
I need to estimate first, second and eventually third moments for a unknown pdf signal.

the signal is 12 bit digitized
the proccess BW is 300KHz

The signal has poisson additive noise (the S/N ratio is a problem when S goes down)


I want to put the math in a FPGA. I need to know the best suitable method for this task. I believe I can't do it in the frecuency field but I'm not sure of that.

Which algorithm should I use?
Which bibliography you recommend me?.

Thanks! :)
 

Re: power spectrum estimation in nongaussian noise

Hi,
check out HOS papers and Book by Nikias and Petropoulu for an introduction. Periodogram is good only for Gaussian Noise. The algorithms are normally available in IEEE papers.

The algorithms are very computation intensive. So try to get to periodogram approach.

B R M
 

power spectrum estimation in nongaussian noise

if periodogram is not working, try autoregressive
 

Re: power spectrum estimation in nongaussian noise

I need just higher order moments not the spectrum, so rxx , rxy, rxxxit's enough.

Non Gaussian process & periodogram: I understand periodogram is only necesary to get the spectrum.

The AR way: Panda, I have to model the process and then get the Spectrum... a bad deal because I don't need the spectrum.
 

Hi,
If you dont need the spectrum and you need only the moments, you can calculate it using some slice algorithms. You may find it in 5-6yrs old ICASSP papers. Currently I dont have access to these.
B R M
 

brmadhukar said:
Hi,
If you dont need the spectrum and you need only the moments, you can calculate it using some slice algorithms. You may find it in 5-6yrs old ICASSP papers.
Click to expand...

Are they so complex algorithms that would be difficult to program a FPGA to do some of it?

May be they are targeted for DSP chips...
 

Hi,
I remember using HOS for Adaptive SP. It involved significant computations. You will have to have at least a floating point library for things similar to what I did. I think the library is available somewhere over the net.
You may try direct computation of the moments for beginning and check what accuracy you require. Use Matlab for this.
I shall PM you the reference I had used tomorrow.

Hope this helps.
BRM
 

    Basil

    Basil

    Points: 2
    Helpful Answer Positive Rating
Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top