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

what is "one sided exponential" sequence?

Status
Not open for further replies.
V

vahidkh6222

Full Member level 2
Full Member level 2
Joined
Oct 11, 2005
Messages
137
Helped
6
Reputation
12
Reaction score
0
Trophy points
1,296
Activity points
2,419
hi friends
i need to simulate an i.i.d one sided sequence with zero mean, but when i use exprnd(0,1,1000) in matlab i get all 0s. olease show mw a good reference or any advise for generating such a sequence.
thanks in advance
 

thanks very muchxulfee, but i think i'm looking for a random sequence with exponential pdf...And the problem is, such a sequence with zero mean returns all zeros as i already mentioned.
 

Hi!

You should check your task requirements. It is impossible to obtain nontrivial random value with zero mean and one-sided pdf.
 
well, i know it;s wierd but...here is the passage from original paper:
"wi(k) (i=1,2) are independent, zero mean, i.i.d, one-sided exponential sequences. the normalized third cumulant is 2 in this case"
 

Could you copy-past some text above this phrase?

It is very difficult to understand the meaning "exponential" from one sentance. For example w1(k) and w2(k) may be exponentialy-decayed i.i.d sequences at k>0. Independence in this case is the independence between two sequences but not between consequent samples into one sequence. Or samething like this.
 
this is all the information presentedm about a MIMO equalizer input vectors. and yes, independence is just about two different sequences w1(k) and w2(k) and has nothing to do with seq. generation... believe me, there is nothing more in the paper.
 

I think you found your phrase in the "Identification and deconvolution of multichannel linearnon-Gaussian processes using higher order statistics and inverse filtercriteria" article. If it is right the term "onesided" is concerned to time origin of input signals. These signals are causal i.e. bouth signals strictly equal to zero for negative time indexes. If time is positive we have vector sequence with conventional exponential distribution of vector's components.
 
thanks a lot. yes thats exaxtly it! :eek:O
but still, do we have any "zero mean" exponential sequence?!! maybe i should just subtract the mean from resulting exp seq?!
 

Ye google rulez:D

I can't write exactly because i have no access to that IEEE transactions. But according to abstract it was especially emphasized by authors the non-Gaussian property of exating inputs. So it looks like magnitude-exponential pdf i.e.
p(x) = 0.5*a*exp(-a*|x|)
 

Which is defined only on one side ie. right sided or left sided
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top